The GERDA experiment for the search of 0νββ decay in ^{76}Ge
GERDA Collaboration, K.-H. Ackermann, M. Agostini, M. Allardt, M. Altmann, E. Andreotti, A. M. Bakalyarov, M. Balata, I. Barabanov, M. Barnabe Heider, N. Barros, L. Baudis, C. Bauer, N. Becerici-Schmidt, E. Bellotti, S. Belogurov, S. T. Belyaev, G. Benato, A. Bettini, L. Bezrukov, T. Bode, V. Brudanin, R. Brugnera, D. Budjas, A. Caldwell, C. Cattadori, A. Chernogorov, O. Chkvorets, F. Cossavella, A. D`Andragora, E. V. Demidova, A. Denisov, A. di Vacri, A. Domula, V. Egorov, R. Falkenstein, A. Ferella, K. Freund, F. Froborg, N. Frodyma, A. Gangapshev, A. Garfagnini, J. Gasparro, S. Gazzana, R. Gonzalez de Orduna, P. Grabmayr, V. Gurentsov, K. Gusev, K. K. Guthikonda, W. Hampel, A. Hegai, M. Heisel, S. Hemmer, G. Heusser, W. Hofmann, M. Hult, L. V. Inzhechik, L. Ioannucci, J. Janicsko Csalty, J. Jochum, M. Junker, R. Kankanyan, S. Kianovsky, T. Kihm, J. Kiko, I. V. Kirpichnikov, A. Kirsch, A. Klimenko, M. Knapp, K. T. Knöpfle, O. Kochetov, V. N. Kornoukhov, K. Kröninger, V. Kusminov, M. Laubenstein, A. Lazzaro, V. I. Lebedev, B. Lehnert, D. Lenz, H. Liao, M. Lindner, I. Lippi, J. Liu, X. Liu, A. Lubashevskiy, B. Lubsandorzhiev, A. A. Machado, B. Majorovits, W. Maneschg, G. Marissens, S. Mayer, G. Meierhofer, I. Nemchenok, L. Niedermeier, S. Nisi, J. Oehm, C. O'Shaughnessy, L. Pandola, P. Peiffer, K. Pelczar, et al. (39 additional authors not shown)
EEur. Phys. J. C manuscript No. (will be inserted by the editor)
The
Gerda experiment for the search of ν ββ decay in Ge K.-H. Ackermann , M. Agostini , M. Allardt , M. Altmann ,E. Andreotti , A.M. Bakalyarov , M. Balata , I. Barabanov ,M. Barnab´e Heider , N. Barros , L. Baudis , C. Bauer ,N. Becerici-Schmidt , E. Bellotti , S. Belogurov , S.T. Belyaev ,G. Benato , A. Bettini , L. Bezrukov , T. Bode , V. Brudanin ,R. Brugnera , D. Budj´aˇs , A. Caldwell , C. Cattadori ,A. Chernogorov , O. Chkvorets , F. Cossavella , A. D‘Andragora ,E.V. Demidova , A. Denisov , A. di Vacri , A. Domula , V. Egorov ,R. Falkenstein , A. Ferella , K. Freund , F. Froborg , N. Frodyma ,A. Gangapshev , A. Garfagnini , J. Gasparro , S. Gazzana ,R. Gonzalez de Orduna , P. Grabmayr , V. Gurentsov , K. Gusev ,K.K. Guthikonda , W. Hampel , A. Hegai , M. Heisel , S. Hemmer ,G. Heusser , W. Hofmann , M. Hult , L.V. Inzhechik , L. Ioannucci ,J. Janicsk´o Cs´alty , J. Jochum , M. Junker , R. Kankanyan ,S. Kianovsky , T. Kihm , J. Kiko , I.V. Kirpichnikov , A. Kirsch ,A. Klimenko , M. Knapp , K.T. Kn¨opfle , O. Kochetov ,V.N. Kornoukhov , K. Kr¨oninger , V. Kusminov , M. Laubenstein ,A. Lazzaro , V.I. Lebedev , B. Lehnert , D. Lenz , H. Liao ,M. Lindner , I. Lippi , J. Liu , X. Liu , A. Lubashevskiy ,B. Lubsandorzhiev , A.A. Machado , B. Majorovits , W. Maneschg ,G. Marissens , S. Mayer , G. Meierhofer , I. Nemchenok ,L. Niedermeier , S. Nisi , J. Oehm , C. O’Shaughnessy , L. Pandola ,P. Peiffer , K. Pelczar , A. Pullia , S. Riboldi , F. Ritter , C. RossiAlvarez , C. Sada , M. Salathe , C. Schmitt , S. Sch¨onert ,J. Schreiner , J. Schubert , O. Schulz , U. Schwan , B. Schwingenheuer ,H. Seitz , E. Shevchik , M. Shirchenko , H. Simgen , A. Smolnikov ,L. Stanco , F. Stelzer , H. Strecker , M. Tarka , U. Trunk , C.A. Ur ,A.A. Vasenko , S. Vogt , O. Volynets , K. von Sturm , V. Wagner ,M. Walter , A. Wegmann , M. Wojcik , E. Yanovich , P. Zavarise ,I. Zhitnikov , S.V. Zhukov , D. Zinatulina , K. Zuber , G. Zuzel INFN Laboratori Nazionali del Gran Sasso, LNGS, Assergi, Italy Institute of Physics, Jagiellonian University, Cracow, Poland Institut f¨ur Kern- und Teilchenphysik, Technische Universit¨at Dresden, Dresden, Germany Joint Institute for Nuclear Research, Dubna, Russia Institute for Reference Materials and Measurements, Geel, Belgium Max Planck Institut f¨ur Kernphysik, Heidelberg, Germany Dipartimento di Fisica, Universit`a Milano Bicocca, Milano, Italy INFN Milano Bicocca, Milano, Italy Dipartimento di Fisica, Universit`a degli Studi di Milano e INFN Milano, Milano, Italy Institute for Nuclear Research of the Russian Academy of Sciences, Moscow, Russia Institute for Theoretical and Experimental Physics, Moscow, Russia National Research Centre “Kurchatov Institute”, Moscow, Russia Max-Planck-Institut f¨ur Physik, M¨unchen, Germany Physik Department and Excellence Cluster Universe, Technische Universit¨at M¨unchen, Germany Dipartimento di Fisica e Astronomia dell‘Universit`a di Padova, Padova, Italy INFN Padova, Padova, Italy Shanghai Jiaotong University, Shanghai, China Physikalisches Institut, Eberhard Karls Universit¨at T¨ubingen, T¨ubingen, Germany Physik Institut der Universit¨at Z¨urich, Z¨urich, SwitzerlandReceived: date / Accepted: date a r X i v : . [ phy s i c s . i n s - d e t ] D ec Abstract
The
Gerda collaboration is performing asearch for neutrinoless double beta decay of Ge withthe eponymous detector. The experiment has been in-stalled and commissioned at the Laboratori Nazionalidel Gran Sasso and has started operation in Novem-ber 2011. The design, construction and first operationalresults are described, along with detailed informationfrom the R&D phase.
Keywords neutrinoless double beta decay · germa-nium detectors · enriched Ge · Cherenkov muonveto
PACS β decay; double β decay; electron andmuon capture · ≤ A ≤ · γ -ray spectroscopy · · The
Gerda experiment (GERmanium Detector Ar-ray [1]) is a search for the neutrinoless double beta(0 νββ ) decay of Ge. The observation of such a de-cay would prove that lepton number is not conserved,and that the neutrino has a Majorana component [2]. Adiscovery of 0 νββ decay would have significant implica-tions on particle physics and other fields, including cos-mology [3]. The importance of the topic has stimulatedthe development of several experimental approaches tothe search for 0 νββ decay on a number of isotopeswhich undergo double beta decay. For recent reviewson the state of knowledge concerning double beta decay a deceased b Present Address:
CEGEP St-Hyacinthe,Qu´ebec, Canada c Present Address:
Laurentian University, Sudbury, Canada d Present Address:
Brookhaven National Laboratory, Upton(NY), USA e Present Address:
Department of Neurosciences and Imaging,University “G. dAnnunzio” di Chieti-Pescara, Italy f Present Address:
Nat. Physical Laboratory, Teddigton, UK g now in private business h Present Address:
Moscow Institute of Physics and Technol-ogy, Russia i Present Address:
II. Physikalisches Institut, U. G¨ottingen,Germany, and Department Physik, U. Siegen, Germany j Present Address:
Kavli IPMU, University of Tokyo, Japan k Present Address:
T ¨UV-S ¨UD, M¨unchen, Germany l Present Address:
Karlsruhe Institute of Technology (KIT),Karlsruhe, Germany m Present Address:
Robert Bosch GmbH, Reutlingen, Ger-many n Present Address:
Photon-Science Detector Group, DESY o Present Address:
University of L’Aquila, Dipartimento diFisica, L’Aquila, Italy p Corresponding Author , email: [email protected] and on running or planned experiments, see Refs. [4,5,6,7,8].The experimental signature for 0 νββ decay is a linein the summed electron energy spectrum appearing atthe Q -value for the reaction, Q ββ . The experimentalresult is a measurement of, or a limit on, the half life, T / , for the process. Within the three neutrino modeland assuming the existence of a significant Majoranacomponent a positive observation of 0 νββ decay wouldpossibly give access to the neutrino mass hierarchy aswell as information on the absolute values of the neu-trino masses. The latter is only possible with knowledgeof the nuclear matrix elements, M ν , as discussed inRef. [9,10,11,12,13]. The statements on the mass alsorequire an understanding of whether the 0 νββ processis solely due to the Majorana nature of the neutrino, orwhether additional new physics processes beyond theStandard Model contribute. A recent review of the par-ticle physics implications of a discovery of 0 νββ decaywas given in Ref. [14].Nuclides that are potentially 0 νββ emitters will de-cay via the Standard Model allowed 2 νββ decay. Bothreactions are second order weak interactions, and there-fore have extremely long half lives. Values have been di-rectly measured for 2 νββ decay in about ten cases andthese are in the range 10 –10 yr [5]. The half lives for0 νββ decay, assuming the process exists, are expectedto be substantially longer. Consequently, 0 νββ decayexperiments must be sensitive to just a few events peryear for a source with a mass of tens to hundreds ofkilograms. Backgrounds must typically be reduced tothe level of one event per year in the region of inter-est (ROI), an energy interval of the order of the energyresolution around Q ββ .Experiments looking for 0 νββ decay of Ge operategermanium diodes normally made from enriched mate-rial, i.e. the number of Ge nuclei, the isotopic frac-tion f , is enlarged from 7.8 % to 86 % or higher. Inthese type of experiments, the source is equal to thedetector which yields high detection efficiency. Addi-tional advantages of this technique are the superior en-ergy resolution of 0.2 % at Q ββ =2039 keV comparedto other searches with different isotopes and the highradiopurity of the crystal growing procedure. Disad-vantages are the relatively low Q ββ value since back-grounds typically fall with energy and the relative dif-ficulty to scale to larger mass compared to e.g. experi-ments using liquids and gases. There is a considerablehistory to the use of Ge for the search for 0 νββ decay.After initial experiments [15], the Heidelberg-Moscow(
HdM ) collaboration [16] and
Igex [17] were the driv-ing forces in this field setting the most stringent lim-its. In 2004 a subgroup of the
HdM collaboration [18] claimed a 4 σ significance for the observation of 0 νββ de-cay with a best value of T / =1.19 · yr; the quoted3 σ range is (0 . − . · yr. To scrutinize this re-sult, and to push the sensitivity to much higher levels,two new Ge experiments have been initiated:
Majo-rana [19] and
Gerda [1]. The latter has been built inthe INFN Laboratori Nazionali del Gran Sasso (LNGS)at a depth of 3500 m w.e. (water equivalent). Whereas
Majorana further refines the background reductiontechniques in the traditional approach of operating ger-manium detectors in vacuum,
Gerda submerses barehigh-purity germanium detectors enriched in Ge intoliquid argon (LAr) following a suggestion by Ref. [20];LAr serves simultaneously as a shield against externalradioactivity and as cooling medium. Phase I of theexperiment is currently taking data and will continueuntil a sensitivity is reached which corresponds to anexposure of 15 kg · yr with a background index (BI) of10 − cts/(keV · kg · yr) [1]). This will be sufficient to makea strong statement on the existence of 0 νββ decay in Ge for the best value given in Ref. [18]. Phase II of
Gerda is planned to acquire an exposure of 100 kg · yrat a BI of 10 − cts/(keV · kg · yr). For pure Majoranaexchange and the case that no signal is seen, this willconstrain the effective neutrino mass (cid:104) m ββ (cid:105) to less thanabout 100 meV with the precise value depending on thechoice of matrix elements [21].The Gerda experiment is described in detail in thefollowing sections. An overview of experimental con-straints and the design is presented first. This is fol-lowed by a description of the Ge detectors. Then, theexperimental setup, electronic readout, data acquisition(DAQ) and data processing are described. As
Gerda
Phase I has been fully commissioned and has starteddata production, the main characteristics of its perfor-mance are given in the final section.
The experimental challenge is to have nearly backgroundfree conditions in the ROI around Q ββ . Typically, back-ground levels are quoted in units of counts per keV perkilogram per year, cts/(keV · kg · yr), since the numberof background events roughly scales with the detectormass, energy resolution and running time. Defining ∆ as the width of the ROI where a signal is searched for,the expected background is the BI multiplied by ∆ inkeV and the exposure in kg · yr. Gerda has set the goalto keep the expected background below 1 event. For ∆ = 5 keV and exposures mentioned above, this impliesa BI of 0.01 and 0.001 cts/(keV · kg · yr), respectively, forthe two phases of Gerda . The main feature of the
Gerda design is to oper-ate bare Ge detectors made out of material enriched in Ge ( enr
Ge) in LAr. This design concept evolved froma proposal to operate Ge detectors in liquid nitrogen(LN ) [20]. It allows for a significant reduction in thecladding material around the diodes and the accompa-nying radiation sources as compared to traditional Geexperiments. Furthermore, the background producedby interactions of cosmic rays is lower than for the tra-ditional concepts of HdM , Igex or Majorana due tothe lower Z of the shielding material. Other backgroundsources include neutrons and gammas from the decaysin the rock of the underground laboratory, radioactivityin support materials, radioactive elements in the cryo-genic liquid (intrinsic, such as Ar and Ar, as well asexternally introduced, such as radon) as well as inter-nal backgrounds in the Ge diodes. These backgroundswere considered in the design and construction phase of
Gerda and resulted in specific design choices, selectionof materials used and also in how detectors were han-dled.Natural Ge ( nat
Ge) contains about 7.8% Ge, andcould in principle be used directly for a 0 νββ decay ex-periment. Indeed, the first searches for 0 νββ decay usednatural Ge detectors [15]. Enriched detectors allow fora better signal-to-background ratio and also yield re-duced costs for a fixed mass of Ge in the experiment.The improvement in signal-to-background ratio origi-nates from two sources: ( i ) many background sources,such as backgrounds from external gamma rays, are ex-pected to scale with the total mass of the detector; and( ii ) the cosmogenic production of Ge and Co in theGe diodes occurs at a higher rate for nat
Ge than for enr
Ge. The lower overall cost is due to the fact that thehigh cost of enrichment is more than offset by the costof producing the extra crystals and diodes required for nat
Ge detectors.Fig. 1 shows a model of the realized design: the coreof the experiment is an array of germanium diodes sus-pended in strings into a cryostat filled with LAr. TheLAr serves both as cooling medium and shield. Thecryostat is a steel vessel with a copper lining used pri-marily to reduce the gamma radiation from the steelvessel. The cryostat is placed in a large water tank,that fulfills the functions of shielding the inner vol-umes from radiation sources within the hall, such asneutrons, as well as providing a sensitive medium fora muon veto system. A similar experimental setup hasbeen proposed previously in Ref. [22]. The detectorsare lowered into the LAr volume using a lock systemlocated in a clean room on top of the water tank. Afurther muon veto system is placed on top of the cleanroom in order to shield the neck region of the cryostat.
Fig. 1
Artists view (Ge array not to scale) of the
Gerda experiment as described in detail in the following sections:the germanium detector array (1), the LAr cryostat (2) withits internal copper shield (3) and the surrounding water tank(4) housing the Cherenkov muon veto, the
Gerda buildingwith the superstructure supporting the clean room (5) andthe lock (6, design modified). Various laboratories behind thestaircase include the water plant and a radon monitor, controlrooms, cryogenic infrastructure and the electronics for themuon veto.
These installations are supported by a steel superstruc-ture. All components are described in the subsequentsections.2.1 Auxiliary installationsThe
Gerda detector laboratory (GDL), located under-ground at LNGS, has been used for R&D for
Gerda as well as for auxiliary tests. It is a grey room equippedwith a clean bench, a glove box and wet chemistry foretching. Ge diodes submerged in LN or LAr can becharacterized in a clean environment without any ex-posure to air. The detector handling described in sec. 3and now adopted for Gerda was developed in GDL.The Liquid Argon Germanium (
LArGe ) appara-tus was installed inside GDL to investigate propertiesof LAr, such as the scintillation light output. It is usedfor studies of background suppression in germanium de-tectors by observing the coincident scintillation light ofthe liquid argon [23] and to exploit the LAr scintillationlight pulse shape properties to recognize the interactingparticle [24].
LArGe is a 1 m low-background cop-per cryostat with a shield consisting of (from outsideto inside) 20 cm polyethylene, 23 cm steel, 10 cm leadand 15 cm copper. The inner walls are covered witha reflector foil with a wavelength shifter coating. Theshifted light is detected by nine 8” ETL 9357 photo-multiplier tubes (PMTs) from Electron Tubes Limited (ETL) [25]. Calibration sources ( Th,
Ra, Co,
Cs) have been placed in- and outside of the cryostatand the event rejection by pulse shape discriminationand scintillation light detection were studied [26]. As aconsequence of these measurements
Gerda decided toimplement a LAr scintillation light veto for Phase II.
LArGe has also been used to understand the back-ground coming from the decay of Ar.In addition to GDL, screening facilities at LNGS, inparticular GeMPI [27] and Gator [28], have been usedextensively. Additional screening facilities have beenused at different locations, including Heidelberg, Geel,and Baksan.Finally, many of the institutes in the
Gerda collab-oration have laboratories which have been extensivelyused in R&D and testing related to the experiment.2.2 Monte Carlo simulationsA full Monte Carlo simulation of the
Gerda experi-ment and of many of the related R&D facilities has beensetup in the form of a general and flexible frameworkbased on
Geant4 [29,30], which is called
MaGe [31].
MaGe has been widely used for
Gerda -related simula-tions and background studies. Conversely, most of theexperimental test stands provided experimental datathat were used to validate and benchmark
MaGe . Adetailed simulation of the
LArGe setup is also avail-able within
MaGe .A few specific
Gerda -related simulations were runusing other codes than
MaGe . In particular, a dedi-cated simulation code was developed to estimate theresidual background in the detector array due to exter-nal γ -rays, produced either in the surrounding rocksor in the cryostat volume [32]. The simulation codeSHIELD [33] was used to optimize the shielding re-quired for the transportation of GeO enriched in Gefrom the enrichment plant to the underground stor-age site [34]. Neutron spectra and fluxes produced by α s from the Th calibration sources via the ( α ,n)reactions were calculated through the SOURCES-4Acode [35]. This section describes the germanium detectors thatrepresent the core of the
Gerda experiment. For Phase Iall eight detectors from the former
HdM and
Igex ex-periments [16,17] were refurbished and redeployed. ForPhase II new material amounting to 50 kg enr
GeO and34 kg of dep GeO was purchased. The dep Ge, materialdepleted in Ge below 0.6 %, was used to check the p + n + groove[ 58 , 80 ] [ , ] Fig. 2
Schematic drawing of a enr
Ge diode currently oper-ated in
Gerda . The ranges of dimensions for the eight detec-tors are given in units of mm. The masses range from 0.98 to2.9 kg. supply chain and methods of Phase II diode produc-tion [36]. The production and characterization of thenew detectors is ongoing.Phase I detectors are based on standard p-type HPGedetector technology from Canberra Semiconductor NV,Olen [37]. Standard closed-end coaxial detectors have a“wrap around” n + conductive lithium layer ( ∼ + contactby a groove; the groove region is usually passivated.The detector geometry for one of the enriched detec-tors is shown schematically in Fig. 2. In normal DCcoupled readout, the p + surface ( ∼ µ m) is connectedto a charge sensitive amplifier and the n + surface is bi-ased with up to +4600 V. In the alternative readoutscheme with AC coupling, the n + contact is groundedand the p + contact biased with negative high voltage(HV). The analog signal is still read out from the p + contact but coupled with a HV capacitor to the ampli-fier.Operation of bare HPGe detectors in cryogenic liq-uids is a non-standard technique. The success of Gerda depends strongly on the long-term stability of the Gedetectors operated in LAr.3.1 Prototype detector testing in LAr and in LN Before deploying the enriched detectors in LAr, bare nat
Ge detectors built with the same technology as thePhase I detectors were used for tests in GDL. A long-term study of the leakage current (LC) of bare detectorsoperated in LN and LAr under varying γ -irradiationconditions was performed. Irradiation of a first proto-type detector in LAr with γ ’s resulted in a continuousincrease of the LC (see Fig. 3, left).The ionizing radiation created the expected bulkcurrent in the detector ( ∼
40 pA), observed as a stepat the start of the γ -irradiation at t ∼ t ∼ ∆ LC ≈
30 pA). No increase ofthe LC was observed with the same detector assemblyin LN after one week of irradiation.The process is reversible as the LC was partly re-stored by irradiation with the same source but withoutapplying HV; the LC was completely restored to itsinitial value by warming up the detector in methanolbaths. These measurements are the first observation of γ -radiation induced leakage current increase for detec-tors of this design operated in this way. The γ -radiationinduced LC was measured for different HV bias values,source-detector configurations and HV polarities [38,39]. Measurements with three prototype detectors usingdifferent sizes of groove passivation (large area, reducedand none) were performed. It was found that reducingthe size of the passivation layer strongly suppresses the γ -radiation induced LC (see Fig. 3, right). The mostlikely explanation is that the LC increase is induced bythe collection and trapping of charges produced by theionization of LAr on the passivated surface of the detec-tor. No γ -radiation induced LC increase was observedwith the prototype without passivation layer.For all stability measurements [39], the detectorswere biased above their nominal operation voltage. TheLC, continuously monitored with high accuracy, was ata few tens of pA for each detector, similar to the valuesmeasured at the detector manufacturer. Detectors withno passivation layer showed the best performance inLAr. Consequently, all Gerda
Phase I detectors werereprocessed without the evaporation of a passivationlayer. Our positive results on the long-term stability ofGe detectors in LAr and LN contradict the statementsmade in Ref. [40].3.2 Phase I detectorsThe enriched Phase I detectors ANG 1-5 from the HdM and RG 1-3 from the
Igex collaborations were origi-nally produced by ORTEC. In addition, six detectorsmade of nat
Ge are available from the GENIUS-TF ex-periment [41]. They have been stored underground andtherefore their intrinsic activity is low. Thus, they havebeen used in the commissioning phase of
Gerda . De-tails of the characterization of the enriched detectorsbefore they were dismounted from vacuum cryostats in2006 are reported in Ref. [42].The Phase I detectors, enr
Ge and nat
Ge, were mod-ified at Canberra, Olen [37], in the period from 2006to 2008. The detector ANG 1 had a previous repro-cessing at the same manufacturer in 1991. The work l ea k age c u rr en t ( p A ) l ea k age c u rr en t ( p A ) LAr LN γ irradiation Prototype 3 (without passivation layer)Prototype 2 (reduced passivation layer)Prototype 1 (full passivation layer)
Fig. 3
Left: γ -radiation induced leakage current (LC) of the first prototype operated in LAr. Right: γ -radiation induced LCfor 3 prototype detectors with different passivation layers. was performed according to the standard manufacturertechnology, however the passivation layer on the groovewas omitted. Leakage current and capacitance of eachdetector were measured in LN at the manufacturer siteafter the reprocessing [39].The detector dimensions after the reprocessing, theoperating bias determined in the LAr test bench ofGDL and with the abundance of Ge measured ear-lier are reported in Table 1. A total of ∼
300 g wasremoved from the detectors during reprocessing result-ing in 17.7 kg enriched diodes for Phase I. The activemasses of the detectors were assessed at typically ∼
87 %by comparing γ -ray detection efficiencies to Monte Carlosimulations of the diodes with dead layer thicknessesvaried [39]. This assessment will be refined with in-situ Gerda data.Cosmogenically produced isotopes Ge and Cocan lead to an internal contamination that represents abackground in the region of interest. The detectors arealways stored at an underground facility to avoid ex-posure to cosmic rays. This applies also for the repro-cessing steps, where the detectors were stored under-ground at the HADES facility [43], located at a depthof about 500 m w.e. at a distance of 15 km from the de-tector manufacturer. The total exposure above groundwas minimized to ∼ Co is on average about(1 − · − cts/(keV · kg · yr). The bulk of the Coactivity comes from the production before the under-ground installation of the detectors for the
HdM and
Igex experiments. The contribution from Ge is neg-ligible since it decayed away.The mounting scheme of the detectors has compet-ing requirements. It must have a low mass to minimize sources of radiation near to the detectors. However, theconstruction must be sufficiently sturdy to provide safesuspension. It must support the cables for detector biasand readout. Furthermore, the diodes must remain elec-trically isolated from all other materials. The chosensupport design is depicted in Fig. 4 where the con-
Fig. 4
Drawing of a Phase I detector assembly. The signalcontact is realized by a conical copper piece (“Chinese hat”)that is pushed by a silicon spring onto the p + contact (in-set left top). High voltage is applied to the n + contact bya copper strip (not shown) pressed by a copper disc whichin turn is electrically insulated by a PTFE cylinder (insetbottom left). The force to achieve good electrical contact isactuated through a copper screw. Masses and dimensions ofthe assembly are given for the RG3 detector. Table 1
Characteristics of the Phase I enriched and natural detectors. The isotopic abundances for Ge, f , of the ANG-type detectors are taken from Ref. [44]; those for RG-type detectors are from Ref. [45]; the natural abundance [46] is taken forGTF detectors. The numbers in parentheses in the last column give the 1 σ -uncertainties (for details see Table 2).detector serial nr. diam. length total operat. abundancename ORTEC (mm) (mm) mass (g) bias (V) f ANG 1 (cid:63) ) 58.5 68 958 3200 0.859 (13)ANG 2 P40239A 80 107 2833 3500 0.866 (25)ANG 3 P40270A 78 93 2391 3200 0.883 (26)ANG 4 P40368A 75 100 2372 3200 0.863 (13)ANG 5 P40496A 78.5 105 2746 1800 0.856 (13)RG 1 † ) 28005-S 77.5 84 2110 4600 0.8551 (10)RG 2 † ) 28006-S 77.5 84 2166 4500 0.8551 (10)RG 3 † ) 28007-S 79 81 2087 3300 0.8551 (10)GTF 32 P41032A 89 71 2321 3500 0.078 (1)GTF 42 P41042A 85 82.5 2467 3000 0.078 (1)GTF 44 P41044A 84 84 2465 3500 0.078 (1)GTF 45 P41045A 87 75 2312 4000 0.078 (1)GTF 110 P41110A 84 105 3046 3000 0.078 (1)GTF 112 P41112A 85 100 2965 3000 0.078 (1) (cid:63) ) produced by Canberra, serial nr. b 89002. † ) as different types of measurements vary, an uncertainty of 2 % is taken in evaluations. tacting scheme is shown as well. In order to reach thebackground goals of Gerda , the amount of materialis minimized. Only selected high radiopurity materialswere used: copper ( ∼
80 g), PTFE ( ∼
10 g), and silicon( ∼ γ ray spectroscopy measure-ments (see sec. 6), combined with Monte Carlo simula-tions give an upper limit on the BI contribution fromthe detector support of ≤ − cts/(keV · kg · yr).One of the prototype detectors was mounted in asupport of the Phase I design to test the electrical andmechanical performance. This confirmed the mountingprocedure, the mechanical stability, the signal and HVcontact quality, and the spectroscopic performance ofthis design. During this test, the energy resolution wasthe same as was achieved previously when the samedetector was mounted in a standard vacuum cryostat,i.e. ∼ Co.Fig. 5 shows one of the Phase I detectors before andafter mounting in its custom made support structure.The Phase I detectors were mounted in their final low-mass supports in 2008 and their performance parame-ters (leakage current, counting efficiency, energy resolu-tion) were measured in LAr as a function of bias volt-age [39]. The detector handling was performed in GDLentirely within an environment of N gas. The LC of themajority of the detectors was at the same level as mea-sured at the detector manufacturer after reprocessing.The detectors ANG 1, ANG 3 and RG 3 showed highLCs even after successive thermal cycling and requiredadditional reprocessing to reach an acceptable perfor-mance. Spectroscopic measurements were performed, as described in Ref. [47], with the preamplifier mounted ina gaseous Ar environment in the neck of the LAr cryo-stat. The energy resolutions of the Phase I detectorswas between 2.5 and 5.1 keV (FWHM) for the 1332 keVspectral line of Co. An improvement of the energy res-olution of the detectors was observed after polishing thediode surface in the location of the HV contact.Since November 2011 all the enriched Phase I de-tectors have been inserted into the
Gerda cryostat.
Fig. 5
Left: A Phase I detector after reprocessing at Can-berra, Olen. The conductive lithium layer (n + contact) andthe boron implanted bore hole (p + contact) are separated bya groove. Right: The detector is mounted upside down in alow-mass holder (groove no longer visible). Table 2
The relative number of nuclei for the different isotopes is shown for the different detector batches. The isotopiccomposition of the depleted material is the average of measurements by the collaboration and ECP; that for natural germaniumis given for comparison. germanium isotopedetector batch Ref. 70 72 73 74 76natural [46] 0.204(2) 0.273(3) 0.078(1) 0.367(2) 0.078(1)
HdM – ANG 1 [55] 0.0031(2) 0.0046(19) 0.0025(8) 0.131(24) 0.859(29)
Igex [45] 0.0044(1) 0.0060(1) 0.0016(1) 0.1329(1) 0.8551(10)
Gerda depleted 0.225(2) 0.301(3) 0.083(1) 0.390(5) 0.006(2)
Gerda
Phase II (cid:63) ) [48] 0.0002(1) 0.0007(3) 0.0016(2) 0.124(4) 0.874(5)
Majorana [56] 0.00006 0.00011 0.0003 0.0865 0.914 (cid:63) ) numbers in brackets represent the range of measurements from ECP. for Phase II of
Gerda . A brief description of the ac-tivities is given here.A batch of 37 . enr Ge was procured by theElectrochemical Plant (ECP) in Zhelenogorsk, Russia [48]in 2005. The isotopic content of the enriched germa-nium is given in Table 2. The enrichment was performedby centrifugal separation of GeF gas, and the enr Gewas delivered in the form of 50 kg enr
GeO .A major concern during all steps is the productionof long-lived radioisotopes via cosmogenic activation, inparticular Ge and Co. Specially designed containerswere used to transport the material [34] by truck fromSiberia to Germany; the enr
GeO was then kept in theHADES facility in underground storage while not beingprocessed.A series of reduction and purification tests with dep Gewas organized. A complete test of the production chainfrom enrichment to the tests of working diodes was per-formed within a year. Based on results on isotopic di-lution and yield, it was decided to further process thematerial at PPM Pure metal GmbH [49]. The process-ing of the enr
GeO took place in spring 2010. The stepsincluded a reduction of GeO to “metallic” Ge, withtypical purity of 3N (99 . ≥ . . . . Co line.Tests in LN and LAr are underway. Five of them havebeen placed into a string and inserted into the Gerda cryostat in July 2011.
Gerda occupies an area of 10.5 × in Hall Aof Lngs between the TIR tunnel and the LVD exper-iment. A model of the experiment is shown in Fig. 1.The floor area has been refurbished with reinforced con-crete for enhanced integral stability and was sealed withepoxy for water tightness. A grid surrounding the watertank is connected to the new
Lngs water collection sys-tem. The various components were erected sequentially.The construction of the bottom plate of the water tank(sec. 4.2) was followed by the installation of the cryo-stat (sec. 4.1) which arrived by a flat-bed truck fromthe manufacturer in March 2008. After the acceptancetests, the water tank construction was resumed and fin-ished in June 2008. Subsequently the
Gerda building(sec. 4.3) was built and on top of it the clean room(sec. 4.4) was erected; the latter houses the lock systemwith a glove box, the calibration system (sec. 4.5) aswell as auxiliary cabinets. The earthquake tolerance ofthe setup was verified by calculating the relative mo-tions of cryostat, water tank and
Gerda building for aseismic event with strength and frequency parametersprovided by
Lngs . The muon veto system (sec. 4.6)consists of two parts, the water Cherenkov detectorwhich is mounted within the water tank and an arrayof plastic scintillators which are located on the roof ofthe clean room.
Gerda cryostat holds 64 m of LAr which servesas medium for the cryogenic operation of the bare Gediodes as well as a shield against the remnants of theexternal γ background penetrating the surrounding wa-ter and against the radioactivity of the cryostat itself.Leakage of radon from the atmosphere into the cryo-stat is prevented by the exclusive use of metal seals inthe joints and valves and by keeping an overpressureof about 3 · Pa against atmosphere. In the originaldesign copper of low radioactivity, i.e. < µ Bq/kg of
Th, was foreseen as production material. However,safety issues and an unexpected cost increase forcedthe change to a stainless steel cryostat with an inter-nal copper shield. Taking into account the measuredradioactivity values of the stainless steel material [57](see sec. 6), the thickness of the copper shield was de-termined by analytical calculations and MC simulationssuch that sources of γ radiation external to the cryostatand the cryostat itself contribute to the BI by about0 . · − cts/(keV · kg · yr) [32].This section describes the cryostat and the cryo-genic system required for its stable operation and someperformance features of the setup. At the end specialsafety aspects are discussed that result from the oper-ation of a cryostat immersed into a large water volumelocated in an underground site. The cross section of the super-insulated cryostat is shownin Fig. 6. The cryostat is supported by a skirt (item 1)at a height of about 1.3 m above the bottom of thewater tank. Access to the volume below the cryostatwithin the skirt is provided by two manholes (item 2).The cryostat consists of two coaxial vessels compris-ing of torospherical heads of 4200 and 4000 mm outerdiameter and corresponding cylindrical shells of about4 m height. The inner vessel rests on eight Torlon [58]pads (item 3) located on the bottom head of the outervessel. Both vessels have a cylindrical neck of 1.7 mheight and are connected at the top. The compensa-tion for thermal shrinkage of the inner container is pro-vided by a double-walled stainless steel bellow in itsneck (item 7). In the upper region the outer neck car-ries four DN200 flanges (item 9) which are protectedagainst the water by a kind of “balcony” surroundingthe neck (item 8). A flexible rubber fabric closes thegap between the water tank roof and the balcony. Theflanges allow access to the volume between inner andouter vessel and they carry the pump and the pressuresensors for the insulation vacuum as well as a safety
662 335 5712 1110 989 4 1313 131 . m . m Fig. 6
Cross section of the LAr cryostat inside the watertank (right part cut away). The following components areindicated: skirt (1), access hole (2), Torlon support pads (3),radon shroud (4), internal copper shield (5), lower and upperheat exchanger (6), bellow in neck of inner vessel (7), balcony(8), DN200 ports (9), manifold (10), bellow between cryostatand lock (11) and DN630 shutter (12). The skirt provides 6mounts for PMTs (13). disc as protection against overpressure. The neck of theinner vessel with an inner diameter of 800 mm providesthe only access to the interior of the cryostat. A mani-fold (item 10) on top of the neck carries the flanges forthe feedthroughs of all devices that penetrate into thecold volume including a filling tube, gas exhaust tube,tubes for active cooling, and feedthroughs for the cryo-stat instrumentation. The Ge diodes are lowered intothe cryostat through a lock which resides in the cleanroom above the manifold (see sec. 4.4). Relative move-ments between manifold and lock are decoupled fromeach other with a flexible bellow of 600 mm diameter(item 11). A DN630 UHV shutter (item 12) on top ofthe bellow allows the stand-alone operation of the cryo-stat without lock.The internal copper shield (item 5) consists of sixty3 cm thick overlapping plates of high purity oxygen freeradiopure (OFRP) copper with a total mass of 16 t. They are mounted on a support ring achieving a copperthickness of 6 cm for the central 2 m high ring (centeredat 4 m height) and of 3 cm thickness in a range of 40 cmabove and below.Radon can emanate from the vessel walls and maybe transported by convection close to the Ge diodes.To prevent this a central volume of about 3 m heightand 750 mm diameter is separated from the rest by acylinder (item 4) made out of 30 µ m thick copper foil.This cylinder is called the radon shroud.During production and after its deployment at Lngs the cryostat has been subjected to several acceptanceand performance tests. Both the inner and the outervessel passed the pressure vessel tests according to theEuropean pressure vessel code PED 97/23/EC. Heliumleak tests for the inner and the outer vessel showed noleak at the 10 − (Pa · (cid:96) )/s range. Evaporation tests withLN established the specified thermal loss of <
300 Wboth at the factory and after delivery. The
Rn em-anation rate of the inner volume of the cryostat hasbeen measured at room temperature at several stageswith the MoREx system [59] (for details see Table 4in sec. 6.2). After iterated cleaning the empty cryostatexhibited the excellent value of (14 ±
4) mBq whichincreased after the mounting of the Cu shield and thecryogenic instrumentation by about 20 mBq at eachstep, leading to a final value of (54 . ± .
5) mBq. A uni-form distribution of this amount of
Rn in the LArwould correspond to a BI ∼ · − cts/(keV · kg · yr). De-pending on its tightness, the radon shroud is expectedto reduce this contribution by up to a factor of seven. The cryogenic infrastructure consists of storage tanks,super-insulated piping, and the systems for vacuum in-sulation, active cooling, process control, and exhaustgas heating. The power for the entire system is takenfrom a dedicated line which is backed-up by the
Lngs diesel rotary uninterruptible power supply.The storage tanks for LN and LAr, about 6 m each, are located at about 30 m distance. To minimizeargon losses they are connected by a triaxial super-insulated pipe (LAr, LN and vacuum super-insulationfrom inside to outside) to the cryostat. The LAr tankhas been selected for low radon emanation. The tankhas been used for the filling of the cryostat and willbe used further for optional refillings. The LAr passesthrough a LN -cooled filter filled with synthetic char-coal [60] to retain radon as well as through two PTFEfilters with 50 nm pore size to retain particles. For thefirst filling the charcoal filter was bypassed. The insulation vacuum has to be maintained in avolume of about 8 m . Out-gassing materials in this vol-ume include about 75 m of multilayer insulation and50 m of additional thermal insulation (Makrolon [61]of 6 mm thickness). A pressure of 10 − Pa was reachedafter two months of pumping with a turbo pump of550 (cid:96) /s pumping speed and intermediate purging withdry nitrogen. After cool down the pressure dropped toabout 2 · − Pa. At a residual out-gassing rate in therange of 10 − (Pa · (cid:96) )/s, the turbo pump is kept runningcontinuously.The active cooling system uses LN as cooling medium.It has been designed [62] to subcool the main LAr vol-ume in order to minimize microphonic noise in the cryo-stat while maintaining a constant (adjustable) work-ing pressure without evaporation losses. This is accom-plished by two LN /LAr heat exchangers (item 6 inFig. 6), spirals of copper tube located in the main vol-ume and at the liquid/gas surface in the neck, respec-tively. With the nitrogen gas pressure of 1 . · Paabsolute, corresponding to a LN boiling temperatureof 79.6 K, the LAr is cooled to about 88.8 K. Since thetemperature is slightly higher than the boiling pointat standard atmospheric pressure, the cryostat builds aslight overpressure until an equilibrium is reached suchthat no argon is lost. The daily LN consumption isabout 280 (cid:96) .In case of an incident like the loss of insulation vac-uum, LAr will evaporate at an estimated rate of upto 4.5 kg/s. The cold gas has to be heated to a tem-perature above 0 ◦ C before it is discharged to the
Lngs ventilation system. This is achieved by a water-gas heatexchanger (see Fig. 8) using the
Lngs cooling water orthe
Gerda water tank reservoir.Complete control over almost all processes is achievedwith a programmable logic controller (PLC) Simatic S7from Siemens which continuously monitors the infor-mation provided by more than 10 redundant pairs ofPt100 temperature sensors distributed in the cryostatvolume, the vacuum gauges, and the level and pressuresensors. To improve the safety further pressure regula-tion was installed, that is independent of the PLC. Theoutput of a stand-alone pressure gauge (SMAR LD301,[63]) regulates directly the positioner of a valve. Twosuch systems are implemented to further increase thereliability. All status information is communicated tothe general
Gerda slow control system (sec. 5.5) andcan be accessed globally via a web-based graphical userinterface that also allows restricted remote control.Since its filling with LAr in December 2009, thecryostat has remained at LAr temperature and oper-ations have been stable. Except for a small refill of LArduring the tuning of the active cooling system and one more following a forced Ar evaporation for a radon mea-surement in the exhaust gas, no additional LAr refillwas necessary. The additional risks of operating a cryostat within awater tank due to the huge latent water heat were ana-lyzed early in the design phase. Specific mitigation mea-sures were realized in the design, construction and theoperation of the cryostat and cryogenic system. Themost important ones are summarized below.The cryostat was designed and produced accordingto the European pressure vessel code for a nominal over-pressure of 1 . · Pa, even though it is operated belowthe limit of 0 . · Pa above which this code applies.An additional safety margin is provided by an increaseof the wall thickness of the cold vessel by 3 mm. The riskfor any leak in one of the vessel’s walls is further reducedby the lack of any penetrations in the inner or outer ves-sel below the water fill level, the 100 % X-raying of thewelds and an earthquake tolerance of 0.6 g. The useof ductile construction materials guarantees the cryo-stat to follow the leak-before-break principle. In caseof a leak, the implementation of a passive insulation atthe outside of the inner and the outer vessel will limitthe evaporation rate to a tolerable maximum of about4.5 kg/s.The oxygen fraction in air is monitored continuouslyfor any low level employing several units placed in the
Gerda building and in the clean room. Further en-hanced safety features include full redundancy of pres-sure and level sensors as well as the use of both a rup-ture disk and a safety valve for overpressure protection.The insulation vacuum is continuously monitored witha residual gas analyzer reading the partial pressures forwater, argon, and nitrogen. This information will beused for diagnostics in case of an unexpected rise in to-tal pressure. In case of a relevant leak the PLC wouldautomatically start the drainage of the water tank. Arealistic test has established the complete drainage tobe possible within less than two hours (see sec. 4.2).4.2 The water tank and its water plantThe water tank when filled with water provides a 3 mthick water buffer around the cryostat whose purposeis fourfold: ( i ) to moderate and absorb neutrons, ( ii )to attenuate the flux of external γ radiation, ( iii ) toserve as Cherenkov medium for the detection of muonscrossing the experiment, and ( iv ) to provide a back-upfor the Lngs cooling water which in case of emergencymight be needed to heat the argon exhaust gas.
The water tank with a nominal capacity of 590 m wasdesigned following the API 650 regulation and accord-ing to the Eurocode 8 for the design of structures forearthquake resistance. It was built completely on siteafter the installation of the cryostat on the pre-installedbutt-welded ground plate (Fig. 7). It consists of a cylin-der of 10 m diameter and 8.3 m height covered bya conically-shaped roof which extends up to 8.9 m;the water level is kept at 8.5 m. AISI 304L stainlesssteel was used exclusively as construction material. Thesheet metal plates for the cylindrical shell have a thick-ness from 7 mm to 5 mm and are joined by butt weldsusing externally (internally) MIG (TIG) welding. Anadditional bottom reinforcement has been applied atthe 1 foot level. Following the UNI EN 1435 code, asignificant fraction of the 400 m length of welds wasX-ray tested.Access into the water tank for the installation andmaintenance of the muon veto (sec. 4.6) is possiblethrough a manhole at the bottom of 1400 ×
800 mm size. The roof has a central hole of 1200 mm diame-ter through which the neck of the cryostat sticks out.The gap between neck and the roof is closed by a flexi-ble membrane made of rubber to block radon and lightfrom the water volume. Radon intrusion is further re-duced by a slightly over-pressurized nitrogen blanketbetween water and roof. Besides numerous small flanges, Fig. 7
The water tank under construction in Hall A of
Lngs in front of the LVD detector. The inset shows how the tank isassembled from top to bottom. The hall crane lifts the upperpart to which another cylinder segment of about 2 m heightis welded. The cryostat in the center is protected by a blackfoamed plastic during the construction of the water tank.2 the water tank has a further DN600 manhole as wellas a DN600 chimney for the PMT cables on the roof,and, at the bottom, two DN300 flanges for fast waterdrainage.The water tank was filled via a dedicated pipelinefrom the Borexino plant [64] with ultrapure water of re-sistivity close to the physical limit of 0.18 M Ω · m. Thestatic test of the water tank consisted in the measure-ment of its radial deformation of the tank as function ofthe water column height and finally applying an over-pressure of 10 Pa. Radial deformations were measuredin three azimuthal locations at a height of 1 m and inone location at a height of 4 m. The maximum deforma-tion was 7 to 8 mm as measured both in the azimuth ofthe manhole at 1 m height and on the opposite side ofthe tank at 4 m height. The deformations were provento be elastic.The water tank exhibits various features to ensuresafe operation (see Fig. 8). A pressure relief valve willopen when the nominal overpressure of (2 − · Pa isexceeded. Complete drainage of the water was demon-strated in less than two hours. A constant drainage ratethrough a new DN250 pipe underneath the TIR tunnelof up to 65 (cid:96) /s is controlled by the PLC. According tothe actual water level, the PLC sets the opening angleof a butterfly valve on that pipe to control the rate.A second pipe, with a maximum flow rate of 16 (cid:96) /s,leads via the grid to the Hall A pits that are devoted tocollecting any fluid accidentally discharged by the ex-periments. In an emergency, a third channel is openedto pump water from the water tank at a rate of 20 (cid:96) /sthrough the heater for the LAr exhaust gas (Fig. 8).This third channel also drains to the pits in Hall A.During such an emergency event, an additional safetyvalve opens a vent to prevent a collapse of the watertank.
The water plant (Fig. 8) has the function to keep thefraction of ions normally existing in the water, espe-cially U, Th, K, as low as possible (fractions of ppm).Also the level of the Total Organic Carbon (TOC) mustbe controlled, otherwise they would cause a gradualdegradation in the optical transparency of the waterover time.The water in the
Gerda tank is kept in constantcirculation by a loop pump at typically 3 m /h. In itsreturn path the water is purified by an “Ultra-Q” unit.This is a special device equipped with four disposablecartridges containing specific resins, that removes TOCand ions (both anionic and cationic) from the water. Fi-nally, the water is filtered for the removal of suspended particles and returned to the bottom of the water tankvia a circular distribution system. The quality of thewater is monitored after the filter by its resistivity andis typically higher than 0.17 M Ω · m. The high light yieldobserved in the muon veto system (sec. 4.6) is furtherproof of the excellent water quality.4.3 The Gerda buildingThe
Gerda building evolved from the need of a super-structure that supports a platform above the water tankand cryostat to host a clean room with the lock systemfor the insertion of the Ge diodes into the cryostat. Theblue beams of the superstructure are visible surround-ing the water tank in Fig. 1. The gap between the watertank and
Lvd is occupied by laboratory rooms on threelevels plus a platform and a staircase. The ground floorhouses the water plant and a radon monitor, the firstfloor two control rooms (one of them dedicated to
Lvd )and the second floor part of the cryogenic infrastruc-ture including the heater for the Ar exhaust gas, safety
Fig. 8
Schematic of the
Gerda water system including thedrainage, the argon exhaust gas heater and the water plant.3 valves and PLC as well as the electronics for the muonveto.4.4 The clean room, twin lock and detector suspensionsystemsThe platform on top of the
Gerda building supportsthe infrastructure for the clean handling and deploy-ment of the Ge detectors into the cryostat without ex-posing them to air. This infrastructure is designed asa gradient of radon reduction and cleanliness (Fig. 9).First a clean room is the working environment for ex-perimenters within which a nitrogen flushed glove boxis the working environment for the detectors. At thecenter a lock system provides a clean change betweenthe environments of the glove box and the cryostat fordetector insertion. The personnel lock and two smallside rooms complete this complex. clean room personnellockpumpsaccess platformm m . m S NWE
Fig. 9
Plan of the platform on top of the
Gerda buildingshowing the clean room and auxiliary cabinets. The positionsof the water tank (1), the cryostat (2) and its neck (3), allbelow the platform, are indicated. The two arms of the lockand detector suspension system (4,5) are connected to thecryostat. The lock is enclosed by a glove box (6). The heightof the clean room varies from 1.3 to 3.6 m.
The clean room is a class 7 room (ISO 14644-1 [65])corresponding to less than 10.000 particles/ft of di-ameter ≥ µ m. An overpressure of up to 30 Pa ismaintained by pressing filtered air into the clean room.The air volume of the clean room can be exchanged 49times per hour. Access to the clean room is via a person-nel lock where an overpressure of 15 Pa is maintained.The temperature inside the clean room is kept constantwith variations of up to ± . ±
20) %. The constancy of these parameters dependsto some extent on the LNGS cooling water supply of theunderground laboratory.The ceiling of the clean room follows the curvedshape of the ceiling of Hall A, such that the central partof the clean room has a height of 3.6 m while the heightat the wings reduces to a maximum of 2.5 m (see Fig. 9).The central part is equipped with two cranes at a heightof 3.3 m that are movable along the south-north (S-N) direction, each with a maximum load of 500 kg.Both the southern wall and the central roof componentare demountable. A maximum load of 150 kg/m canbe supported on the roof, greater than the load of theplastic muon veto system (sec. 4.6). Adjacent to theclean room is an electronic cabinet with a cable trayfeedthrough to the clean room. Another adjacent roomhouses the pumps for the gas system of the lock.The class 7 specifications have been met during alltimes while the clean room was operating. A LabViewprogram monitors and outputs in a web interface thefollowing parameters: particle measurements, radon con-tent, overpressure, temperatures, and humidity. The twin lock system for Phase I consists of two inde-pendent arms (Figs. 9 and 10) that are connected withthe cryostat via a cluster flange on top of the DN630shutter (bottom inset of Fig. 10, see also sec. 4.1 andFig. 6). Inside each arm is a cable chain (top inset ofFig. 10), the mechanics for lowering the detector stringsinto the cryostat and lights and cameras for observationduring this procedure. One lock arm supports three de-tector strings inside a vertical tube of 250 mm diam-eter, while the other supports a single detector stringinside a vertical tube of 160 mm diameter. Since thearms are part of the argon gas volume during data tak-ing, they are built according to the European pressurevessel code. The locks are constructed from stainlesssteel tubes that are connected either by welding or byCF metal seals. The vertical section where the detectorstrings are mounted are both located inside a glove boxwhere HEPA filters further reduce particle concentra-tion. Each vertical part consists of two about 1 m longtubes which exhibit the functionality of an independentlock for one or three detector strings, respectively.Each lock arm may be closed from the cryostat byindividual shutters (item e in Fig. 10) allowing for theindependent operation of each when the DN630 shutteris opened. The removal of the lower part of the verti-cal tube (item d) allows for the insertion of detector ca b dea c: fixation of cable chain and cable feedthroughecluster flange s eDN630 shutter HEPA filter mm Fig. 10
Sketch of the twin lock for Phase I with its two arms on top of the DN630 shutter flange. The transparent bluearea indicates the glove box with the HEPA filters (red). Each arm has an individual lock shutter (e) above which the verticaltube (d) can be removed to allow the insertion of the detector strings. The upper inset demonstrates the principle of the locksystem: steel band (red) and cable chain with cables (black), winch (a), linear pulley (b), fixation of cable chain with cablefeedthroughs (c), movable tube (d), and individual lock shutter (e). The inset at the bottom right details the DN630 shutter,the cluster flange, the individual tube shutters and also two of the three source insertion systems (s) above the DN40 shutters.A picture of the bottom side of the cluster flange is shown in the bottom left. Visible are the DN160 and DN250 openings aswell as the 3 smaller openings for the sources with the tantalum absorbers (and 2 spare holes). string(s) into the lock. The Ge diodes are transferred inevacuated containers into the glove box that is purgedwith boil-off nitrogen gas. Within the glove box, germa-nium diodes and their front end electronics are assem-bled into strings of up to three diodes each (discussedin sec. 5.1; a fully mounted string is shown in Fig. 17).These strings are then transferred into the lock. Afterthe closure of the lock, it is evacuated and purged withargon gas. The two lock volumes are connected individ-ually to a pumping station and to the cryostat througha dedicated gas system. The latter has been helium leaktested at a level of 10 − (Pa · (cid:96) )/s.As radon can diffuse through plastic, metal sealsare used almost exclusively for the lock system. Allnon-metal materials were screened for radon emana-tion (see sec. 6, Table 5). The DN630 shutter is con- nected with Helicoflex metal seals, while a Kalrez sealis employed for the shutter itself. The flange with themotor axle feedthrough has a double seal EPHD O-ring. To avoid radon diffusion through this non-metalseal, the feedthrough is constantly pumped. The leakrate of the motor connection was measured to be about10 − (Pa · (cid:96) )/s.The scheme of the suspension system is shown in thetop inset of Fig. 10. The cable chain is fixed inside thelock (item c) and runs along the 3.6 m long horizontaltube. It is deflected at the far end of the tube by 180 ◦ around the “linear pulley” (item b), a pulley that is freeto move in the horizontal direction by sliding on a linearbearing. Above the cryostat the chain is deflected by90 ◦ vertically. The linear pulley is connected to a metalband that rolls around a winch (item a) fixed to the Table 3
Cables deployed in the 1-string and 3-string locks.cable ref. type 1-string 3-stringHabia SM50 [66] 50 Ω , coaxial 15 24SAMI RG178 [67] HV (4 kV), coaxial 4 -Teledyne Reynolds 167-2896 [68] HV (18 kV), coaxial - 10Teledyne Reynolds 167-2896 [68] HV (5 kV), unshielded 1 2total number 20 38 axle of a stepper motor. By unrolling the metal band,the linear pulley moves towards the cryostat neck andthe chain can be lowered into the cryostat.The cable chain supports the detectors mechanicallyand provides a conduit for the signal and high voltagecables to operate them. It is constructed from stainlesssteel that was selected for radiopurity. Its cross sectionis 21 ×
13 mm with a fillable area of 17 × . Table 3shows the configuration of the respective cable bundlesfor the 1- and 3-detector string case. In the 1-stringbundle all cables are wrapped in a PTFE spiral coiledtube. This protects them against damage while mov-ing inside the cable chain during its operations. Thehigher number of cables needed to operate nine detec-tors could be accommodated only by weaving the cableswith PTFE thread into flat cables and protecting themagainst friction with the bottom of the cable chain bya thin metal band (see Fig. 11). Fig. 11
Woven cable bundles in the cable chain of the 3-string lock.
The chain movement and the shutter status are con-trolled by a dedicated PLC. Inductive sensors are usedas end switches. The position of the chain is determinedredundantly by counting the number of turns of the mo-tor and by a measuring tape with holes. An optical sys-tem counts evenly spaced holes in a steel tape that is un-rolled as the chain is lowered. A friction clutch mountedbetween feedthrough and motor gear protects againstexcessive force transmission onto the cable chain. 4.5 The calibration systemRegular calibration measurements with radioactive γ sources provide the data necessary to determine theenergy calibrations and resolutions of the diodes and tomonitor their stability. The energy scale is tracked viamonitoring of specific γ lines to identify periods in timefor which single diodes showed a degraded performance.These time periods can be identified and omitted orspecially treated in the final analysis.In order to calibrate the detectors within the LArcryostat, three Th calibration sources are broughtinto the vicinity of the crystals. This is achieved bythree vacuum sealed mechanical systems (Fig. 12) thatare mounted on top of the cluster flange (Fig. 10). Thesystems can be individually decoupled from the cryo-stat via DN40 gate valves with electrical state indica-tors. To ensure that the background from the calibra-tion sources is negligible during physics data taking,the sources are mounted on top of tantalum absorbersof 60 mm height and 32 mm diameter (Fig. 12). Thesemovable absorbers rest inside the ones mounted on thecluster flange (left inset of Fig. 10). Each absorber withits source is connected to a perforated stainless steelband which is deflected by 90 ◦ before being rolled on Fig. 12
A schematic view of one of three units of the cali-bration system (see also insets of Fig. 10 and Ref. [69]).6 spindles mounted behind horizontal band guides. Thespindles are connected via magnetofluid sealed rotaryfeedthroughs to planetary geared DC motors. Frictionclutches between the feedthroughs and the motors pro-tect against excessive force transmission on the steelband. The sources are moved with a speed of 10 mm/s.Each unit has two redundant positioning systems.An incremental encoder counts the holes of the steelband that is perforated at 4 mm pitch. The incremen-tal encoder is mounted below the 90 ◦ deflection pointat the end of the band guide. At the same position amicroswitch defining the null position is mounted. Acustom designed feedthrough, mounted on a CF flange,passes electronics for the incremental encoder as well asa gas line with a VCR 1/2” gas connection. The latterallows to pump and purge the source gas volume afterthe installation and before the shutter to the cryostatis opened.The second positioning system is based on a mag-netic sampling multi-turn absolute encoder with 13 bitresolution, registering changes in position even if notpowered. The absolute encoder is mechanically coupledvia a gear wheel to the external drive shaft on whichthe DC motor is mounted.The three source systems are controlled by a com-mon control unit enabling the communication betweena micro-controller and a PC via an RS422 interface.Each calibration source can be individually moved viaa control panel displaying the actual position and sta-tus of the respective unit. The control panel also al-lows for a manual movement of the sources. Correctionfunctions for the thermal contraction of the steel bandimmersed in the cryo-liquid are applied when calculat-ing the position. The incremental encoder serves as themain positioning system, while the absolute encoder iscalibrated with respect to it.The RS422 interface allows to remotely control thesystem via a LabView GUI [70]. The GUI allows to au-tomatize source movements, to change relevant settings,and to monitor the status of the sources and the control-ling unit. A closed or undefined gate valve state vetoesany motor activity on the LabView side. Malfunctionsof the systems are monitored by the micro-controllerthat blocks any further movement of the sources in casean error occurs.Tests of the calibration systems prior to mountingon the cryostat showed an accuracy of the incrementalencoder of ± ± ± Gerda cryostat in June 2011. During the commissioning of a prototype sys-tem, a source dropped to the bottom of the cryostatdue to the rupture of the supporting steel band. ForPhase I, the resulting contribution to the BI is negli-gible ( ≤ . · − cts/(keV · kg · yr) based on ∼
22 kBqin November 2011). The final version of the calibrationsystem is working without any problems.The energy calibration of the diodes is performed byusing 7 prominent lines in the
Th spectrum: 510.8 keV,583.2 keV, 727.3 keV, 860.6 keV, 1620.5 keV, 2103.5 keVand 2614.5 keV. For the calibration function a secondorder polynomial is used to account for ballistic defectsof the measured pulses and for non-linearities of theelectronics. Calibration spectra with the resolutions ofall detectors are shown in sec. 7.2.4.6 The Muon VetoThe Gran Sasso overburden of 3500 m w.e. reduces theflux of cosmic muons to about 1.2 /(h · m ) and shiftsthe mean energy to 270 GeV. Muons penetrating thedetector will lose energy by both electromagnetic inter-actions and by inelastic reactions with nuclei in whichhigh energy neutrons can be produced. These neutronswill cause inelastic interactions themselves and producemore isotopes and neutrons. Hence muons are both adirect and indirect background source.The instrumentation of the water buffer surround-ing the cryostat provides a cost-effective solution forthe identification of muons by the detection of theirCherenkov light with PMTs. Muons that enter the cryo-stat through the neck might only pass through a smallwater volume. An array of plastic scintillators on theroof of the clean room provides additional covering todetect muons passing this region. Signals from both de-tector systems are combined as a muon veto serving thegermanium DAQ. The muon veto system is designed toreduce the BI contribution from the direct muon eventsto a level of 10 − cts/(keV · kg · yr) at Q ββ in the regionof interest. MC simulations have been used to optimize the setupand in particular to define the number of photon detec-tors inside the water tank [71]. A reflective foil glued onthe walls of the water tank and cryostat contributes sig-nificantly to the light collection efficiency. This VM2000foil, produced by 3M [72], has a reflectivity of >
99 %over a wavelength range of 400 to 775 nm and per-forms well as wavelength shifter for UV light that is re-emitted in the visible spectrum. The foil has a rathersmall thickness of 206 µ m, i.e. ∼ . Almost all outer surfaces of the cryostat, the inner wall and thefloor of the water tank are covered with this foil.For the Cherenkov light detection 8” PMTs fromETL, type 9350KB/9354KB [25], have been selected.They withstand an overpressure of 2 · Pa absolutewhich is more than the pressure due to the maximumwater height of 8.5 m. Since the PMTs are located out-side the cryostat, there are no stringent constraints ontheir radioactivity. Nevertheless, 23 low-activity PMTsavailable were mounted in the almost closed water vol-ume within the skirt of the cryostat (see Figs. 1 and 6)and on the bottom plate of the water tank. The selectedETL 9454KB PMTs have e.g. a potassium content re-duced from 300 ppm to 60 ppm. All PMTs are encapsu-lated in stainless steel housings to prevent water fromreaching the electrical contacts as shown in Fig. 13. In
Fig. 13
Schematic drawing of the encapsulation for thePMTs of the Cherenkov muon veto. The PMT, the oil andthe silicone are not shown. addition, the housing also acts as a mechanical supportand mount for the PMTs. It consists of a steel cone,fixed to a bottom plate. To keep the PMT in position,the electrical base is fixed with polyurethane. As addi-tional waterproofing, the electrical contacts are pottedwith silicone. The optical face of the housing is closedwith a polyethylene terphthalate (PET) window. Thevolume between window and PMT is filled with mineraloil for a better matching of the refractive indices.The electrical power for a PMT and its signal read-out is provided by a single HV coaxial RG 213C/Ucable with polyurethane cladding that is designed forunderwater applications [73]. To facilitate timing, allcables have the same length of 35 m. In the electronicsroom, a splitter separates the HV and the signal lines.Extensive tests have been performed to secure the un-derwater tightness of the capsules. One prototype wasoperated at full HV inside a pressure tank for several years without degraded performance. Independent longterm tests of material degradation and cable perfor-mance were also performed.The water tank is equipped with 66 PMTs yieldinga nominal coverage of 0.5%. Of these, 6 are mountedon the skirt facing inwards into the volume below thecryostat. Due to the few small openings (see. Fig. 6),this part is an almost independent water volume sepa-rated from the main. In the main volume, sets of 8 and12 PMTs are mounted to the floor of the water tankfacing upwards in rings of 5.5 m and 8.5 m diameter,respectively. The remaining 40 PMTs are mounted tothe outer wall of the water tank facing inwards in fourrings of 10 PMTs at the heights of 2 m, 3.5 m, 5 m,and 6.5 m. The high voltage is supplied via a CAENSV1527LC crate housing 6 CAEN A1733P high volt-age cards [74] with 12 positive high voltage channelseach.Five diffuser balls [75], each equipped with a singleLED, are distributed throughout the water tank. Whenpulsed they illuminate the entire water tank volume al-lowing for tests of all PMTs simultaneously. The PMTgain is adjusted and the calibration, in units of photoelectrons (p.e.), is made using this system. In addition,each PMT can be triggered individually through an op-tical fiber for calibration, monitoring, and testing [76].The initial HV was set for a gain of 2 · for each PMT. The second part of the muon veto system consists of36 plastic scintillator panels located on the roof of the
Gerda clean room above the neck of the cryostat. Eachscintillator panel consists of a sheet of plastic scintilla-tor UPS-923A [77] with dimensions of 200 × × ,an attached electronics board housing a PMT (one of17 H6780-2 [78] or 19 PMT-85 [79]) and trigger elec-tronics. The light produced inside the plastic panel isguided to the PMT via 56 Y11 [80] optical fibers. Theyare glued to both of the 200 × side areas of thepanel.The 36 panels are stacked in three layers with 12modules each, covering an area of 4 × in the N-Sdirection and centered at the neck of the cryostat. Thepanels in the second layer are placed directly on top ofthe first in the same orientation. The inner 8 modulesof the third layer are rotated 90 ◦ degrees with respectto the lower modules to create a finer pixelization. The data acquisition for the muon detectors is describedin sec. 5.2.2. For a valid trigger of the Cherenkov sys-tem, at least 5 FADCs have to deliver a trigger signal within 60 ns. The threshold in each FADC channel isset such as to accept single photons with about 80%efficiency. The Cherenkov muon veto system is run-ning smoothly since the beginning of the commissioningruns. Three out of the 66 PMTs in the water tank havebeen lost during two years of operation.A standard pulse from the Cherenkov detectors hasa width of about 20 ns followed by a small overshootand an electronic reflection around 350 ns after themain pulse. The heights of each secondary pulse wasless than 1/10 of the main pulse causing no problem forthe trigger system. Pulse height calibration is employedto adjust the gain of the individual PMT. At sufficientlow light pulser rates the PMTs can be set to the sameresponse by adjusting slightly the respective HV. SinceSeptember 2010 the PMTs have been checked period-ically for stability. Only a few HV channels needed tobe re-adjusted during that period. The single photonpeak is clearly distinguishable with a peak-to-valley ra-tio approaching 3.Within an event, the arrival time of pulses with alarge light production is widely spread with differencesup to 340 ns. Nevertheless, around half of the PMTs firewithin the first 60 ns; therefore, this time interval hasbeen chosen as coincidence time window for the trigger.The time spread is produced in part by the reflectionson the VM2000 foil for the benefit of higher light yield.As to the plastic muon veto system, the triple coin-cidence between the layers allows for a clear separationof γ background and muons. This is demonstrated bythe spectra shown in Fig. 14. In the singles spectrum(blue), the low energy part due to γ rays is dominatingand it exhibits a long tail to higher energies. The triplecoincidences reveal unambiguously the minimum biassignal of muons. Fig. 14
Spectra taken with the plastic panels: singles (blue),triple coincidences (pink), and their difference (green).
The triggers of the Cherenkov and the plastic de-tector systems are combined via a logic OR that isrecorded by the germanium DAQ.
Gerda , the same schemewould result in a distance of 10 m between the cold andthe warm part of the CSP. The signal propagation timeto close the feedback loop would consequently be longerthan 100 ns. This would limit the bandwidth or lead tooscillations and the pulse shape information would belargely lost. To avoid this loss we operate the entire CSPin LAr. The minimal allowed distance between the de-tectors and the preamplifier depends on the radioactiv-ity of the latter. The schematic of the implemented CSP(called CC2 [81]) is shown in Fig. 15. The input JFETis a BF862 from NXP Semiconductors and the secondstage is the AD8651 from Analog Devices. Both compo-nents are used in commercial packages. Three channelsare integrated on a single layer Cuflon PCB (deliveredby Polyflon [82]). The feedback and test pulse capac-itors are implemented as stray capacitances betweentraces on the PCB board (see Fig. 16). Tantalum filtercapacitors are used solely and a separate line driver isomitted to limit the radioactivity (see sect. 6.4).The CC2 is located inside a copper box (Fig. 17,right) that provides electromagnetic shielding. The in-put wires connecting to the detectors are copper stripswith 2 × cross section produced by wire erosion
50 50V CC V FET C F R F V EE V ref C T R D V = 2.5 CC V = −2.7
EEref
V = 1.5
FET
V = 13 C = 0.35p F R = 500M F C = 0.35p T R = 4.7k D AD8651BF862detectorHV filterHV
10m 10m 10m 10m
CC2 + − argon testpulse 10m
50 FADCFADC
Fig. 15
The scheme of the
Gerda front end circuit includinggrounding and cable lengths. The parts within the dotted boxare on the CC2 PCB. The red dashed line shows the limitsof the argon volume. The resistor values are given at roomtemperature.9
Fig. 16
The CC2 PCB front and back side integrating threechannels. from screened material. The insulator is a PTFE tube.The same scheme is used for the last part connectingthe high voltage cable to the detector. All copper stripsare fixed along the detector supports to avoid micro-phonics.Realizing the CC2 with the values given in Fig. 15,its specifications are: sensitivity of 180 mV/MeV, in-put range at least 10 MeV, power consumption of lessthan 45 mW/channel, cross talk between channels ofless than 0.1 %, rise time with terminated analog outputof typically 55 ns, decay time of 150 µ s. The noise (con-verted to energy equivalent FWHM for Ge) is typically0.8 keV + 0.024 keV/pF for a 10 µ s semi-Gaussian pulseshape with 8 % systematic uncertainty attached. For a600 g coaxial detector the energy resolution achievedwas 1.96 keV at the 1274 keV γ line of Na.
Fig. 17
Left: a string of three enr
Ge detectors is insertedinto the mini-shroud. This work is performed in the glove boxof the clean room. Right: closed detector string and 3-channelCC2 preamplifier inside a copper box about 30 cm above thestring. The connections between CC2 and detectors are madewith Teflon insulated copper strips that are tightly fixed toprevent microphonics. In the background, part of the 3-stringlock is visible.
A pulser signal is sent periodically to the test pulseinput of the CC2 (Fig. 15). The voltage step at thecapacitor C T injects a fixed charge at the input of theCC2 and thus allows a monitoring of the entire readoutchain during data taking.Thin coaxial cables from Habia (type SM50, 94 pF/m,0.9 Ω /m, [66]) are used for the analog outputs of theCC2 as well as for power supply (see Table 3). WeldedBNC feedthroughs act as seal between air and the cryo-stat/lock. The ground is connected via the lock with thecryostat and water tank. RG178 cables transmit signalsfrom the BNC connectors to the FADC in the electron-ics cabinet where the analog signals are digitized. Thetotal cable length that the CC2 must drive the signalover amounts to 20 m.The HV feedthroughs between air and the argon gasinside the lock are custom made. For leak tightness thebraid of the HV cable is replaced for a few centimetersby a solid wire. The latter together with the solderedconnection and part of the cable is then encapsulatedwith epoxy (Stycast (R) FT2850, [83]) inside a 5 cmlong stainless steel pipe with a CF16 flange at one end.This solution avoids discharges in the argon gas withup to 6 kV bias on the cable. The HV cable shieldingis connected to this pipe and thus also to the lock. Atthe air side of the feedthrough, a filter is mounted toreduce electromagnetic noise.The HV cables inside the lock are Sami RG178 [67]and Teledyne Reynolds 167-2896 [68]. They end about30 cm above the top detector from where the abovementioned copper strips are used. The HV bias is pro-vided by NIM modules from CAEN (N1471H, 4ch PowerSupply, [74]).5.2 The data acquisitionThe data from the Ge detectors and from the muonveto system are acquired with two different data ac-quisition systems. Both systems are synchronized by acommon GPS pulse per second (PPS) signal. All signalsare digitized by FADCs and the energy is reconstructedoffline. The custom made Phase I DAQ [84,85] for the ger-manium readout consists of NIM modules, PCI basedreadout boards and external logic for the trigger genera-tion. Each NIM module digitizes 4 channels. It acceptsboth single-ended and differential signals. The signalpolarity as well as signal attenuation (0 dB/12 dB) orgain (0-6-12-18 dB) of the analog input stage can beselected via jumpers. The offset is adjustable and the bandwidth of the anti-aliasing filter is 30 MHz. Analog-to-digital conversion is based on the Analog DevicesAD6645 A/D converter (14 bits, 100 Msamples/s). Thedigitized data are processed with a trapezoidal filterwith programmable threshold to generate a trigger, av-eraged to monitor the baseline and sent via an LVDSlink to a PCI board. The synchronization of the dif-ferent NIM modules is ensured by a common externalclock and BUSY signal. The latter blocks the incomingdata during readout.The custom-made PCI readout boards are mountedin a personal computer running the Linux operatingsystem and are operated at 32 bit/33 MHz allowing fora maximum transfer rate of 132 MB/s. Selecting a 40 µ swindow, four of the 10 ns samples are added (i.e. thesampling rate is reduced from 100 to 25 MHz) to reducethe data rate and readout time. For pulse shape anal-ysis a 5 µ s trace around the rising edge of the signalis stored in addition at full sampling rate. The lengthallows for a 10 µ s shaping time for the moving windowdeconvolution algorithm[86], and thus no informationfor the energy reconstruction is lost in the compres-sion. A 32 bit trigger counter and a 64 bit 100 MHztimestamp are saved together with the data. NIM logicbuilds the OR of all triggers and generates the BUSYsignal.A Qt based [87] comprehensive graphical user inter-face for the whole system was implemented. A JAVA-based Graphical Analysis tool was developed for onlinemonitoring of the data. Test measurements were per-formed with a BEGe detector. The energy resolutionwas similar to the one obtained with spectroscopy am-plifiers.In addition, a copy of the hardware used for the dig-itization and triggering of the muon veto signals (seebelow) is available for the readout of the germaniumdetectors. Some parameters of the commercial FADCs,e.g. the shaping of the signal for the trigger, are ad-justed to the preamplifier pulse shapes and the tracelength is set to 160 µ s. Both germanium DAQ systemsare operated in parallel. The muon data acquisition is installed in a VME cratehousing 14 FADCs of type SIS 3301 from Struck (8 chan-nels, 100 MHz, 14 bit, [88]). Each card has 2 memorybanks of 128k samples which are divided into 4k sizeper event. If one bank is full, writing continues to thesecond bank while the first one can be read out. Thisreduces the dead time to less than 0.1% during normaldata taking. Each channel is equipped with an analoganti-aliasing filter with 30 MHz bandwidth and with a trapezoidal filter for the digitized data. If the filter out-put is above threshold a trigger is generated. The log-ical OR of all 8 triggers in a FADC module is fed intoa custom made VME board, called MPIC. The MPICgenerates a global trigger if a programmable number ofcards output a trigger within a coincidence time win-dow. This card also provides a time stamp for the eventthat is synchronized to the GPS PPS signal with 10 nsprecision. If a trigger occurs, a “stop pulse” is fannedout to all FADC cards to stop writing to the circularevent buffer such that the data is saved for readout.Upon a trigger, 4 µ s traces for all channels are storedon disk. The stop pulse is also digitized as an additionalanalog input by the germanium DAQ to easily veto co-incidence events. Delayed coincidences can be detectedby comparing the event time stamps between the muonand germanium events in the offline analysis.The muon veto calibration mentioned in sec. 4.6.3 isperformed by powering five LED drivers with a digital-to-analog converter (PAS9817/AO [89]) in connectionwith a CAEN V976 [74] fan-out for pulser signals. Thelight from the LEDs is uniformly distributed to all thePMTs through five diffuser balls.5.3 Data handling The binary raw data format is defined by the differ-ent data acquisition systems. In order to optimize theanalysis streaming and to provide a unique input in-terface for the analysis, all raw data are converted toa common standardized format. MGDO (
Majorana - Gerda
Data Objects) [90] is a software library jointlydeveloped by
Gerda and
Majorana , that containsgeneral-purpose interfaces and analysis tools to supportthe digital processing of experimental or simulated sig-nals. The custom data objects available in the MGDOpackage are used as reference format to store events,waveforms, and other DAQ data (time stamps, flags).The MGDO data objects are stored as
Root files [91].The set of
Root files produced by the conversion ofraw data is named
Tier
Tier Q ββ arenot exported to Tier
Gelatio [92] contains nearlyindependent and customizable modules that are appliedto the input
Tier a new Root file (
Tier
Tier i files can be created that contain additional parame-ters evaluated from more advanced analysis (e.g. cal-ibrated energy spectra). The information of the sameevent stored in different Tier i files can be accessed bymeans of the Root friendship mechanism [91].
The data acquisition systems store data undergroundon a server with 14 TB space. Every night, the newlyaccumulated data are transferred to a
Gerda server inthe LNGS computing center that has 36 TB disk space.This server is only accessible for a small number of userssuch that the raw data are hidden and only blindeddata are available for analysis. Copies of the raw dataare stored at LNGS, in Heidelberg, and Moscow.
The event reconstruction of new data occurs automat-ically once per day. Since our rate is low, it is possibleto store filtered information like the event energy, pulserise time or baseline level in a data base. An interfaceallows simple access with a web browser or, alterna-tively, by a user written C++ program [94]. The eventtraces stored in
Tier
Gerda network structure
Gerda has a dedicated network in Hall A. It is con-nected to the external laboratories above ground by twodedicated multi-modal optical fibers. They connect toa network switch [95] that offers access security andadvanced prioritization and traffic-monitoring capabil-ities. The different network lines are routed inside the
Gerda infrastructure.The switch is directly connected to a dedicated ser-ver [96] that provides network routing facilities andacts as a firewall and user authentication server. Atthe moment, this is the only public service available di-rectly from external networks and it is used to accessall
Gerda internal network resources and services. APort Address Translation (PAT) network device is usedinternally, to translate TCP/UDP communications be-tween
Gerda private network computers and publicnetwork hosts. The following centralized services are available: ( i ) NIS -server for user authentication, ( ii ) DNS -server forhost name resolution, ( iii ) DHCP -server for the DAQ/slowcontrol machines and all the computers attached tem-porarily to the network (i.e. laptops), and ( iv ) Web -server for the whole experiment.In order to provide access to internal
Gerda re-sources (mainly internal Web servers), a proxy servicehas been setup. Thus, it is possible to access internalWeb servers through the main
Gerda
Web server.5.5 The slow controlThe
Gerda slow control system [97] is responsible for:1. monitoring of parameters characterizing the statusof the subsystems (temperature, pressure, detectorcurrents, etc.),2. control and monitoring of low and high voltage powersupplies through a graphical user interface (GUI),3. storage of the monitored values in a database forlater retrieval;4. alarm handling,5. web pages for subsystems breakdown,6. online histograms for the relevant parameters,7. reliable remote monitoring of the whole experiment.The slow control consists of four building blocks. Adatabase is the core of the slow control system. It storesboth the data and the configurations. PostgreSQL [98]was selected as relational database SQL compliant withthe capability of an embedded procedural language (PL).In case of a high number of records in the data ta-bles, the database will be split in two: a so-called onlinedatabase where all the data up to one week are storedand the historical database where older data are copiedregularly after data compression.The acquisition task is performed by a pool of clientseach serving a dedicated hardware subcomponent. Theclients store the acquired data in the database. Depend-ing on the specific hardware, different types of connec-tions are used by the clients: web access, CANbus, serialRS232, etc. All data written into the database have aproper time stamp, that constitutes the main methodto study correlations. All hardware settings are storedin the database.Alarms generated automatically by some compo-nents go directly to the LNGS safety system and tothe
Gerda on call experts and the slow control systemwill record the event into the database for future anal-ysis. The alarm manager is a supervisor process thatretrieves data from the database and is able to generatewarnings or error messages in case of a malfunctioningsensor. The system is completed by a web interface wherealarms, instant and historical data (through histograms)and the status of the clients can be seen. The controlinterface is based on HTML. The data are updated au-tomatically using Ajax [99] in pull manner.The database has been operational since autumn2009. Data collected in two and a half years are only 94MB. This is in part due to the use of a data reductionpolicy at the level of the readout of some subcompo-nents (cryostat, clean room, water loop).
A very careful selection of materials is critical to achieveour goal of one to two orders of magnitude reductionin backgrounds relative to previous experiments. ForPhase I, this selection was carried out by using state-of-the-art screening techniques during the design andconstruction phases. The screening facilities continueto be used in the preparations for Phase II of
Gerda .Material screening was performed mainly with thefollowing three techniques:1. Gamma ray spectroscopy with High Purity Ger-manium spectrometers in four underground labo-ratories: at the Max-Planck-Institut f¨ur Kernphysik(MPIK) in Germany, HADES (IRMM) in Belgium,the Baksan Neutrino Observatory (BNO INR RAS)in Russia and at LNGS in Italy. The ultimate de-tection limit for the best spectrometers in deep un-derground laboratories lies around 10 µ Bq/kg for
Ra and
Th [100,101].2. Gas counting with ultra-low background proportionalcounters. They were originally developed at MPIKfor the
Gallex solar neutrino experiment [102] andare used in
Gerda for
Rn measurements.3. Mass spectrometry with Inductively Coupled PlasmaMass Spectrometers (ICP-MS). The
Gerda collab-oration has access to two ICP-MS machines, one atLNGS and one at INR RAS, Moscow.In addition, some dedicated samples were analyzed withNeutron Activation Analysis (NAA) and Atomic Ab-sorption Spectroscopy (AAS).Altogether almost 250 samples were screened bygamma ray spectroscopy. The main focus was on elec-tronics components (about 85 samples), metal samples(about 65 samples, mostly stainless steel and copper)and plastic materials (about 50 samples). Also the Rnemanation technique was extensively applied (about120 samples) and about 20 samples were screened byICP-MS. In this section some selected results, most rel-evant to the construction, will be given. Some moreresults can be found in Refs. [57,103,104]. 6.1 Argon purityThe
Rn concentration of commercial liquid nitro-gen was measured [59] and its purification to a levelof 1 µ Bq/m at standard temperature and pressure hasbeen demonstrated in the past for Borexino [105]. For
Gerda the same questions arose for liquid argon sincethe
Rn concentration in freshly produced argon wasfound to be in the range of mBq/m (STP) which isabout an order of magnitude higher than for nitrogen.While this is not so relevant for the first cryostat filling,a constant refilling was considered for the case that theactive cooling would fail (see sec. 4.1.1).Argon purification tests based on radon adsorptionon low temperature activated carbon traps were per-formed with gaseous and liquid argon. For the gas phase,reduction factors of more than 1000 were achieved fora 150 g trap [106]. These are similar to the resultsachieved for nitrogen [105]. For the liquid phase, inmost cases a reduction factor of 10 could be achievedfor a small 60 g column. In Gerda an activated car-bon column ( ∼ Rn concentration by two orders of mag-nitude. All measurements were performed with the Mo-bile Radon Extraction unit (MoREx, [105]).6.2 Radiopurity of the cryostatBesides the argon, the second largest mass item in closecontact to the diodes is the cryostat. It is made ofaustenitic stainless steel with an additional inner cop-per shield (see Figs. 1 and 6). The stainless steel wasprocured in more than 10 relatively small batches ofa few tons and roughly a 50 kg sample of each batchwas screened with gamma ray spectrometers [57]. Dur-ing this campaign it was discovered that stainless steelmay have low
Th activity that is about 10 timeslower than what was known from earlier screening cam-paigns [107]. Finally, the cylindrical part of the cryo-stat, closest to the diodes, could be constructed fromstainless steel batches with a
Th concentration below1 mBq/kg. All other batches have a
Th concentra-tion below 5 mBq/kg. Another contamination in stain-less steel is Co. In the batches for
Gerda , a mean Co activity of 19 mBq/kg was found [57]. The avail-ability of low radioactivity stainless steel led to a sig-nificant reduction in the necessary mass of the innercopper shield.Any
Rn released from the inner surface of thecryostat will be dissolved in the liquid argon and maybe transported to the germanium diodes by convection.Therefore, the
Rn emanation rate of the cryostatwas measured after its construction. The cryostat was Table 4
Measurements on
Rn emanation of the
Gerda cryostat at room temperature after various stages of construction.no. date description result [mBq]1 Nov 2007 after construction and first cleaning 23.3 ± ± ± ± ± sealed, evacuated and filled with Rn-free nitrogen gasthat was produced with MoREx. After a certain timein which
Rn could accumulate, the nitrogen was agi-tated (to assure a homogeneous radon distribution) anda sample of a few cubic meters nitrogen was extracted.Then the
Rn concentration in this aliquot was mea-sured with low background proportional counters andthe result was scaled to calculate the
Rn emanationrate of the entire cryostat. The measurement was re-peated after various modifications of the cryostat andthe results are summarized in Table 4.The first two measurements were performed whenthe cryostat was still empty, i.e. just the surface ofthe stainless steel vessel was under investigation. Thecleaning then performed was a pickling and passivationtreatment with an acidic gel. In the first measurementa
Rn emanation rate of 23 mBq was measured. Thisreduced by a second cleaning cycle to a level of about14 mBq. After the copper shield was installed a sub-sequent measurement showed an increase of the
Rnemanation rate by about 20 mBq. A plausible hypoth-esis was that dust was introduced during the coppermounting. However, this explanation was rejected be-cause thorough surface wiping did not improve the re-sult significantly (see measurement No. 4).The final configuration of the cryostat includes amanifold through which all tubing is distributed, a com-pensator to connect it to the lock, a radon shroud (seesec. 3) and many sensors and safety devices. The
Rnemanation rate of the cryostat in its final configurationis (54 . ± .
5) mBq. Assuming a homogeneous distribu-tion of
Rn in the liquid argon, this would result in acontribution to the BI at Q ββ of 7 · − cts/(keV · kg · yr).To reduce this background, a cylinder made from 30 µ mthick copper foil (called radon shroud, see item 4 inFig. 6) was installed around the diodes with the inten-tion that Rn that is emanated from the walls is keptat sufficient distance from the diodes.6.3 Radon emanation of components inside the lockThe lock system is directly connected to the
Gerda cryostat (see Fig. 6 and sec. 4.4.2). Thus,
Rn that is emanated inside the lock may be dissolved in the liq-uid argon and can contribute to the background. Con-sequently, the selection of low-emanating constructionmaterials for the lock and items inside the lock wasa rigorous process. Flanges to the outside were sealedwith metal gaskets whenever possible. At places whereO-rings had to be used Kalrez [108] O-rings were cho-sen to avoid VITON, which is known to be a relativelystrong source of radon. The
Rn emanation rates ofall Kalrez O-rings that are used in the lock system wereinvestigated and it was confirmed that they are muchradiopurer than VITON. As a result of these measure-ments, an upper limit of 0.6 mBq can be given for theintegrated
Rn emanation rate of the subset of O-rings that are in direct contact with the inner volumeof the lock.Table 5 summarizes the results of all the other com-ponents in the lock that were screened for their
Rnemanation rate. As can be seen in the right column theintegrated radon emanation rate of all components isless than 17 mBq. This is low compared to the
Rnemanation rate of the cryostat. Moreover, there are coldcopper surfaces in the argon gas phase just above theliquid level which will act as a getter. Therefore, the
Rn emanation of the lock is a minor source of back-ground for
Gerda .6.4 Further selected screening resultsBefore the construction of
Gerda it was already knownthat high purity copper is one of the most radiopurematerials [101]. Therefore, it was the natural candidatematerial for the construction of the low mass diodeholder (see Table 6). As insulating material, one ofthe most promising candidates from previous measure-ments is PTFE. A batch of extruded PTFE was pur-chased that was produced under particularly clean con-ditions and screened with the Ge
MPI spectrometer [103].Finally, radioactivity measurements of the
Gerda front-end electronics have been performed (see sec. 5.1).Particular efforts have been made to produce a lowradioactivity version of the circuit. Some of the keypoints to achieve these results are: manufacturing of Table 5
Radon emanation of non-metallic materials used in the lock. The amount of the material used and the correspondingemanation is listed. Values indicated by * are estimated by the detection of
Ra using γ -ray screening. They are conservativeupper limits since not all Rn will escape the material.component amount material total Rnemanation rateLED 4 pieces (207 ± µ BqKappa camera 4 pieces < µ Bqinductive end switch 4 pieces mostly steel (73 ± µ Bqmeter drive head 2 pieces (860 ± µ Bqmeter drive plug 2 pieces (400 ± µ Bqpulley bearings 12 pieces Iglidur < µ Bqlinear pulley guides 4 pieces Iglidur < µ BqO-ring seal shutter 1 piece Kalrez (400 ± µ BqO-ring motor feedthrough 2 pieces EPDM (7.8 ± µ BqHV cables SAMI RG178 40 m (300 g) < µ Bqsignal cables Habia SM50 508 m (273 ±
50 ) µ BqLV supply TR 5 kV 62 m (50.4 ± µ BqThermovac pressure gauge [109] 2 pieces < µ BqBD diff. pressure sensor [110] 3 pieces (117 ±
18 ) µ BqHV cables TR 18 kV 1.53 kg (100 m) * < < Table 6
Gamma ray screening results for selected materials. Given are 1 σ -boundaries or 90 % limits. Note, one PCB boardserves three detectors. component amount K Rn Th[mBq/kg] [ µ Bq / kg] [ µ Bq / kg]copper detector support 80 g/det. < . < < . ± .
11 25 ± ± ± ± < . ± . ±
28 150 ± the printed circuit board on a specifically selected low-radioactivity substrate (Cuflon), minimization of thenumber of tantalum decoupling capacitors, integrationof low value capacitors as stray capacitance betweentraces directly on board, and careful selection of pas-sive physical components and soldering paste. To reacha BI < − cts/(keV · kg · yr), the Monte Carlo predictsa maximum allowed activity for the front end electron-ics of 2 mBq for Ra and 500 µ Bq for
Th witha separation of 30 cm between the electronics and thetop detectors. The average measured activity of a set ofthree preamplifiers is (286 ± µ Bq and (150 ± µ Bqin
Ra and
Th, respectively, including the pins.Thus, the radiopurity limits are met for the 5 PCBspresently in use.
The construction of the apparatus was completed inJune 2010. The commissioning phase started with theoperation of refurbished nat
Ge diodes from the GENIUS- TF experiment [41], in order to minimize the poten-tial risks for the enr
Ge detectors. A larger backgroundthan expected at Q ββ and an intense line at 1525 keVwas discovered. The origin and mitigation was studiedin the following months (see sec. 7.3). In June 2011a string of enr Ge diodes was deployed for further pre-liminary tests including various operational configura-tions of the detectors and the electric stray fields. Thecommissioning phase was completed on November 1,2011. All components had met their design specifica-tions and an adequate background index was reached;thus, physics data taking of Phase I was started onNovember 9, 2011. A blinding window of 40 keV widtharound Q ββ is in place since January 11, 2012. The rawdata are written to disk, however events with energiesfrom 2019 to 2059 keV are not exported to the Tier mance of the apparatus is summarized here. Resultsare shown demonstrating that a low background hasactually been reached via thorough material selectionand screening. The stability of the performance of thecomplete Gerda setup at LNGS is inferred from theenergy calibrations and the first spectra. The perfor-mance particulars are obtained on the basis of physicsruns between November 2011 and May 2012, which re-sulted in an exposure of 6.10 kg · yr for the enriched de-tectors and 3.17 kg · yr for the natural detectors. Thesedata are collected with an overall live time (calibrationruns subtracted) of about 90%.7.1 The performance of the muon vetoThe PMTs of the muon veto have been checked for pulseheight stability for more than one year. A satisfying in-dividual stability is reflected in the constant averagelight output per muon event per day (Fig. 18, squaresand right scale). This constancy is mandatory for a re-liable determination of the muon rate that is shown bythe crosses in Fig. 18 (left scale). Except for short termfluctuation the rate is consistent with a 2 % sinusoidalvariation with a period of about one year. This is a well-known phenomenon [111] that will be verified when alonger period of data is available.The observed muon rate in Gerda results in a pre-liminary value of (3.42 ± · − cts/(m s) which com-pares very well with the recent Borexino result of(3.41 ± · − cts/(m s) [111].Fig. 19 shows the multiplicity M , the number ofCherenkov PMTs fired. The spectrum is taken withtrigger signals from both muon veto systems with athreshold of 1 photoelectron (p.e.). The expected lightyield is roughly 200 to 300 photons for every centime-ter traversed by a muon. Since almost all surfaces of thewater tank and cryostat are covered with the VM2000 calender time c oun t s / d ay muon veto stability trigger rateaverage light d ay ) · p . e ./ ( eve n t G E RDA - Fig. 18
The average light output per event and day(squares, right scale) of the Cherenkov muon veto. The dailyrates (crosses, left scale) are rather constant. foil, one would expect that most of the muon eventswill cause a high multiplicity of triggered PMTs. Thelow coverage of 0.5% of the surface by PMTs is compen-sated by the reflectivity and wavelength shifting proper-ties of the VM2000 foil. There is, indeed, a rise towardshigh multiplicities as predicted by the MC simulations,but there is also a prominent enhancement observedin the low multiplicity region below M = 20 which isnot present in MC. The low multiplicity bump around M = 10 vanishes for events triggered by the plastic pan-els only. Therefore, it is unlikely that it is caused di-rectly by muons. The hypothesis of local radioactivitycreating scintillation light in the VM2000 foil is still in-vestigated. Triggers from the water Cherenkov cannotcontribute to M <
5. The increase close to M = 0 origi-nates from triggers of the triple layered plastic scintilla-tor when the muon hits the plastic but misses the water.Increasing the trigger threshold to 30 p.e. (dashed line)removes the intensity at low M .The lower limit of the muon detection efficiency(MDE) is estimated for a threshold of 30 p.e. amountingto (cid:15) md = (97.2 ± (cid:15) md = (99.1 ± > > multiplicity0 10 20 30 40 50 60 no . o f e v en t s ( no t no r m a li z ed ) Cherenkov OR panel eventsall events above 30 p.e.
Cherenkov PMT multiplicity
Fig. 19
Measured multiplicities of all mounted CherenkovPMTs without cut on the number of detected photoelectrons(full line) and with a cut of ≥
30 p.e. (dashed line).6 energy [keV] ANG2 energy [keV] ANG3 energy [keV] ANG4 energy [keV] ANG5 energy [keV] RG1 energy [keV] RG2 energy [keV]
550 600
550 600
550 600
550 600
550 600
550 600 c oun t s GERDA 12-09
Fig. 20
The energy spectra of the six enriched germanium detectors are plotted for a calibration with
Th. The blow-upson the right show the fit results for the 583.2 keV and the 2614.5 keV lines including the values for a Gaussian FWHM. but at least two germanium detectors fired ( α emittersfrom the U/Th decay chain have energy > (cid:15) mr = (97 . +1 . − . ) % (median with 68 %central interval), which is in a good agreement with thesimulations.With the measured efficiency the background in-dex due to un-identified prompt muons is estimated as < − cts/(keV · kg · yr), which is well below the speci-fications needed for Phase I and II [71].7.2 Stability of Ge detectorsInitially, eight detectors from Igex and
HdM have beenin operation in the
Gerda cryostat. Two of them, ANG 1and RG 3, developed high leakage currents at the be-ginning of Phase I. These detectors have been removedfrom the analysis of Phase I data. For some time how-ever, they have been used as veto to suppress multi-siteevents. The remaining total mass for analysis is 14.6 kgwith an average enrichment of 86% in Ge correspond-ing to 165 moles.Energy calibrations are performed on a (bi)weeklybasis with the
Th sources. Spectra of the six ac-tive enriched detectors are shown in Fig. 20, includingscaled subplots for the 583.2 and the 2614.5 keV lines.The high count rates cause pile up that would mani-fest itself in tails on the low energy side of the peaks. Proper pile up rejection algorithms and further dataquality cuts have been applied before the fitting [93].The peaks are fitted well by a Gaussian and an errorfunction representing the background. The results areshown by the red lines and the FWHM of the Gaus-sian is given in keV. Values between 4.2 to 5.3 keV(FWHM) at 2614 keV have been obtained. These canbe translated to a mass weighted average of 4.5 keV(FWHM) at Q ββ =2039.01(5) keV [112]. The resolutionof the 2614.5 keV line for all detectors during the firstmonths of data taking is shown in Fig. 21. No significantvariation or trend is visible for this period.The same is also true for the gain, which normallychanged only after some power cycling or temperaturedrifts. The 2614.5 keV γ -line positions in the calibra-tion spectra are stable in time as shown in Fig. 22 asthey fall into a range of ± Nov/21 Dec/21 Jan/20 Feb/19 Mar/21 Apr/20 May/20 Jun/19 r e s o l u t i on ( F W H M ) [ k e V ] ANG2ANG3ANG4ANG5RG1RG2
Th @ 2614.5 keV energy resolution for
GERDA 12-11
Fig. 21
The energy resolution of the germanium detectors isplotted for several energy calibrations with the
Th source.7
Table 7
The background index deduced (without pulse shape analysis) from the event count in the indicated energy windows ∆ E for different running conditions during the commissioning and the first part of Phase I. Corresponding values are shownalso for the
Igex and
HdM experiments.experiment diodes ∆ E exposure background indexdiode environment (keV) (kg · yr) 10 − cts/(keV · kg · yr) Igex [17]vacuum, Cu enclosed enr 2000-2500 4.7 26
HdM [44]vacuum, Cu enclosed enr 2000-2100 56.7 16
Gerda commissioningLAr nat 1839-2239 0.6 18 ± ± +1 . − . Gerda
Phase ILAr, Cu mini-shroud nat 1839-2239 (cid:63) +1 . − . LAr (diodes AC-coupled) nat 1839-2239 (cid:63) +1 . − . LAr, Cu mini-shroud enr 1939-2139 (cid:63) +0 . − . (cid:63) ) excluding the blinded region of Q ββ ±
20 keV interesting energy at Q ββ . The two lines at ± ± Q ββ . The gainshifts within the ROI thus are typically less than 1 keV.This value is small compared to the average FWHM of4.5 keV and shows that the data from all periods canbe added in the search for the peak of the 0 νββ decay. Nov/21 Dec/21 Jan/20 Feb/19 Mar/21 Apr/20 May/20 Jun/19 s h i ft o f . k e V pea k [ k e V ] -2.0-1.5-1.0-0.50.00.51.01.52.0 + 1.3 keV- 1.3 keV GERDA 12-11
ANG2ANG3ANG4ANG5RG1RG2
Fig. 22
Variations of the 2614.5 keV γ line between succes-sive calibrations. The green lines indicate ± Q ββ if scaled linearly in energy. Gerda
The commissioning of
Gerda started with a string ofthree bare low background nat
Ge diodes, and yieldeda surprisingly large BI on the order of the
HdM and
Igex experiments (18 · − cts/(keV · kg · yr), see Ta-ble 7). As another surprise, the line at 1525 keV from K, the daughter of Ar, appeared in the spectra withan intensity much higher than expected on the basis ofthe upper limit for the ratio Ar/ nat
Ar determinedby V.D. Ashitkov et al. [113]. The published limit of < × − g/g at 90% confidence level converts to anupper limit of 41 µ Bq/kg for Ar. These observationsled to the working hypothesis that charged ions, and inparticular the progeny K, are drifting in the electricfield of the bare Ge diodes that are biased with 3 to4 kV via the large n + surface (see sec. 3 and Fig. 2).The concentration of radioactive impurities near the Gediodes can increase. Further studies with different biasschemes confirmed this hypothesis. A major improve-ment of the BI was achieved by enclosing the string ofdetectors with a cylinder, made of 60 µ m thick Cu foil,called “mini-shroud” (BI ≈ . · − cts/(keV · kg · yr)).The volume of LAr from which the ions can be collectedonto the surface of the detectors is reduced and bulkconvection of the LAr near the detectors is prevented.In fact, operating the Ge diodes in AC-coupled mode( n + surface grounded and p + contact biased) withoutmini-shroud but with adequate shielding of the p + con-tact, i.e. without external electrical stray field, yieldeda similar BI of 6 . · − cts/(keV · kg · yr) (see next tolast line in Table 7). For the Phase I physics run, thehermeticity of the mini-shroud, as well as the shield-ing of the HV cables, was further improved in orderto avoid any leakage of electric field lines into the LArvolume. The improvement with respect to the precur-sor experiments is evident. The stability of the BI mustbe proven for a longer period of time.An analysis of the intensity of the 1525 keV linegives a concentration for Ar that is about twice theliterature limit. This estimate is based on the assump-tion of a homogeneous distribution of this isotope out-side the mini-shroud.The intensity of γ lines was investigated in order toidentify sources of backgrounds. The results are com- Table 8
Counts and rates of background lines for the enriched and natural detectors in
Gerda in comparison to the enricheddetectors of
HdM [42]. Upper limits correspond to 90 % credibility interval. The central value is the mode of the posteriorprobability distribution function and the error bars account for the smallest interval containing 68% probability. nat
Ge (3.17 kg · yr) enr Ge (6.10 kg · yr) HdM (71.7 kg · yr)isotope energy tot/bck rate tot/bck rate rate[keV] [cts] [cts/(kg · yr)] [cts] [cts/(kg · yr)] [cts/(kg · yr)] K 1460.8 85 / 15 21 . +3 . − .
125 / 42 13 . +2 . − . ± Co 1173.2 43 / 38 < . . +2 . − . ± < . < . ± Cs 661.6 46 / 62 < . < . ± Ac 910.8 54 / 38 5 . +2 . − .
294 / 303 < . ± . +3 . − .
247 / 230 2 . +2 . − . ± Tl 583.2 56 / 51 < . < . ± . +1 . − .
10 / 0 1 . +0 . − . ± Pb 352 740 / 630 34 . +12 . − . . +9 . − . ± Bi 609.3 99 / 51 15 . +3 . − .
351 / 311 6 . +3 . − . ± . +3 . − .
194 / 186 < . ± . +1 . − .
24 / 1 3 . +0 . − . ± . +0 . − . . +0 . − . ± piled in Table 8 for the natural and the enriched de-tectors in comparison to numbers from HdM [42]. Therate estimates are based on a Bayesian approach start-ing with a flat prior probability distribution function.The general observation is an achieved reduction byabout a factor of 10 with respect to the
HdM exper-iment. The composition of the background in relationto the screening results will be discussed in a futurepublication.Additional contributions to the BI will result fromradioactive surface contaminations such as
Pb aswell as from cosmogenically produced radioisotopes withinthe diodes. These contributions are expected to be smalland will require large data sets to evaluate.Finally, it is worth to mention that auxiliary exper-iments were performed to study the cross sections ofcosmogenic activation of steel and other constructionalmaterials [114], the inelastic neutron scattering [115],the neutron activation cross sections, and the γ decayspectra [116,117,118]. The deduced contributions to theBI are in the order of few 10 − cts/(keV · kg · yr).First energy spectra for enriched and natural diodesare shown in Fig. 23. Notice, that the spectrum fromthe natural detectors has been renormalized to matchthe exposure of the enriched diodes. The low energypart is dominated by the β decay of Ar which hasan endpoint energy of 565 keV. The well known ac-tivity of A ( Ar)= [1.01 ± ± νββ events in the range from 600 to1400 keV for the enriched detectors is clearly visible.The BI of (2 . +0 . − . ) · − cts/(keV · kg · yr) for theenriched detectors is evaluated in the energy region Q ββ ±
100 keV with the 2019 to 2059 keV window ex-cluded (green bar in Fig. 23). This value is an order ofmagnitude lower than the one for the very same detec-tors in their previous shielding in the
HdM and
Igex experiments (see Table 7). energy (keV)0 500 1000 1500 2000 2500 c oun t s / ( k e V k g y r) -1
10 nat. scaled to exposure of enr. yr · enriched detectors, 6.10 kg yr · natural detectors, 3.17 kg yr · enriched detectors, 6.10 kg yr · natural detectors, 3.17 kg - b Ar K - k e VK - k e V B i - k e V B i - k e V T l - k e V bbn (cid:190)(cid:190) G E RDA - Fig. 23
Spectra taken with enriched (red) and non-enriched(blue) detectors during the same time period. The nat
Gespectrum has been normalized to match the exposure of enr
Ge. The blinding window of Q ββ ±
20 keV is indicated asgreen bar. Identified γ lines are indicated.9 Gerda searches for 0 νββ decay of Ge using a newexperimental concept. Bare germanium diodes are op-erated successfully in a 4 m diameter cryostat filled withLAr, requiring only a small amount of radiopure ma-terials as mechanical and electrical support. Shieldingagainst external background is achieved by LAr and anadditional shell of 3 m of water.The experiment started commissioning in May 2010and in November 2011 with physics data taking (Phase I).The experience gained so far shows that all componentswork well.1. The operation of the cryostat inside the water tankis stable and safe.2. Bare germanium diodes are operated reliably in liq-uid argon over a long time and the implementedhandling procedure ensures that many operationalcycles do not deteriorate the performance.3. The readout electronics is balancing the partiallyconflicting requirements of good energy resolution,low radioactivity, and operation at LAr tempera-ture.4. The water tank instrumentation ensures a high vetoefficiency of muon events and only a tolerable loss of3 out of 66 PMTs have stopped functioning duringa 2 year period.5. Data acquisition and monitoring of the ambient pa-rameters operate reliably.6. The implemented software allows for a fast recon-struction of the data together with a good monitor-ing of data quality.The experience from the (bi-)weekly calibrations showsthat the gain drifts of the entire readout chain are typ-ically smaller than 1 keV at Q ββ . This is small enoughto ensure that adding all data will not result in relevantshifts of peak positions or deteriorations of resolutions.The surprisingly large background from K, the Ar progeny, experienced during the commissioningcan be mitigated by two methods: encapsulation of eachdetector string by a closed thin-walled copper cylinderor AC coupling of the detectors. In both cases the elec-tric field outside of the encapsulation is minimized.The Phase I background is determined currently to(2 . +0 . − . ) · − cts/(keV · kg · yr). This value and the in-tensities of gamma lines show an order of magnitudeimprovement compared to the previous HdM and
Igex experiments. In the absence of a signal and given thecurrent BI,
Gerda expects to set 90 % probabilitylower limits of T / > . · yr for an exposure of20 kg · yr. Acknowledgments
The
Gerda experiment is supported financially by theGerman Federal Ministry for Education and Research(BMBF), the German Research Foundation (DFG) viathe Excellence Cluster Universe, the Italian IstitutoNazionale di Fisica Nucleare (INFN), the Max PlanckSociety (MPG), the Polish National Science Centre (NCN),the Russian Foundation for Basic Research (RFBR),and the Swiss National Science Foundation (SNF). Theinstitutions acknowledge also internal financial support.The
Gerda collaboration thanks the directors andthe staff of the LNGS for their continuous strong sup-port of the
Gerda experiment.Preparing and setting up the infrastructure of the
Gerda experiment was made possible only throughthe indispensable help of R. Adinolfi Falcone, T. Apfel,P. Aprili, J. Baumgart, G. Bucciarelli, M. Castagna,F. Costa, N. D’Ambrosio, S. Flicker, D. Franciotti, H.Fuchs, H. Hess, V. Mallinger, P. Martella, B. M¨ork,D. Orlandi, M. Reissfelder, T. Schwab, R. Sedlmeyr,S. Stallio, R. Tartaglia, E. Tatananni, M. Tobia, D.Wamsler, G. Winkelm¨uller, and T. Weber.The prolific cooperation with P. Vermeulen and J.Verplancke from Canberra SNV, Olen in the contextof refurbishment of the enriched germanium diodes isappreciated.
References
1. I. Abt et al. , (GERDA collaboration),
GERDA:The GERmanium Detector Array for the searchof neutrinoless ββ decays of Ge at LNGS ,Letter of Intent to LNGS, see also: hep-ex/04040390, Proposal to LNGS 38/04 (2004), ; PhD theses: .2. J. Schechter and J.W.F. Valle, Phys. Rev.
D 25 , 2951(1982).3. S. Davidson, E. Nardi and Y. Nir, Phys. Rept. , 105(2008).4. F. Avignone III, S. R. Elliot, J. Engel, Rev. Mod. Phys. , 481 (2008).5. A.S. Barabash, Phys. Atom. Nucl. , 603 (2011).6. J.J. Gomez-Cadenas et al. , Rivista Nuovo Cimento ,29 (2012).7. S. R. Elliot, Mod. Phys. Lett. A 27 , 1230009 (2012).8. B. Schwingenheuer, procs. 12 th International Con-ference on Topics in Astroparticle and UndergroundPhysics, Munich, Sep 5 - 9, 2011.9. M. Doi, T. Kotani and E. Takasugi, Prog. Theor. Phys.Suppl. , 1 (1985).10. J. Suhonen and O. Civitarese, Nucl. Phys. A847 , 207(2010), and refs. therein; J. Suhonen, J. Phys., Conf.Ser. , 042026 (2012).11. J. Menendez, A. Poves, E. Caurier, and F. Nowacki,Nucl. Phys.
A818 , 139 (2009), and refs. therein.012. V. A. Rodin et al. , Nucl. Phys.
A766 , 107 (2006)[Erratum-ibid.
A793 , 213 (2007)]; F. Simkovic et al. ,Phys. Rev.
C 79 , 055501 (2009); V. A. Rodin, J. Phys.,Conf. Ser. , 042025 (2012).13. J. Barea and F. Iachello, Phys. Rev.
C 79 , 044301(2009).14. W. Rodejohann, Int. J. Mod. Phys.
E20 , 1833 (2011).15. E. Fiorini et al. , Phys. Lett.
B 25 , 602 (1967);F.T. Avignone III et al. , Phys. Rev. Lett. , 721(1983);J. J. Simpson et al. , Phys. Rev. Lett. , 141 (1984);E. Bellotti et al. , Phys. Lett. B 146 , 450 (1984);H. Ejiri et al. , Nucl. Phys.
A448 , 271 (1986);D. Caldwell, J. Phys.
G 17 , S137 (1991);P. Fisher et al. , Phys. Lett.
B218 , 257 (1989);A. Morales et al. , J. Phys.
G 17 , S211 (1991);J. Busto et al. , Nucl. Phys.
A513 , 291 (1990);A.A. Vasenko et al. , Mod. Phys. Lett. A5 , 1299 (1990);I.V. Kirpichnikov, Preprint ITEP91-91 (1991).16. HdM -collaboration: H. V. Klapdor-Kleingrothaus et al. ,Eur. Phys. J.
A12 , 147 (2001).17.
Igex -collaboration: C. E. Aalseth et al. , Phys. Rev. D , 092007 (2002);C. E. Aalseth et al. , Phys. Rev. C , 2108 (1999);D. Gonzalez et al. , Nucl. Phys. B procs. , 278 (2000);D. Gonzalez et al. , Nucl. Instrum. Methods A515 , 634(2003)18. H.V. Klapdor-Kleingrothaus et al. , Phys. Lett.
B586 ,198 (2004).19.
Majorana -collaboration: R. Henning, CIPANP 2009Conference Proceedings, AIP Conf. Proc. , 88(2009); J. Wilkerson et al. , J. Phys., Conf. Ser. ,042010 (2012).20. G. Heusser, Ann. Rev. Nucl. Part. Sci. , 543 (1995)..21. A. Smolnikov and P. Grabmayr, Phys. Rev. C81 ,028502 (2010).22. Yu. G. Zdesenko et al. , J. Phys. G: Nucl. Part. Phys. , 2129 (2001).23. M. Di Marco, P. Peiffer and S. Sch¨onert, Nucl. Phys.Proc. Suppl. , 45 (2007).24. P. Peiffer et al. , JINST , P08007 (2008).25. 8” PMT, type 9357, 9350KB & KB9354KB. ET Enter-prises Limited, Uxbridge, UK.26. M. Agostini et al. , J. Phys., Conf. Ser. , 042009(2012).27. H. Neder, G. Heusser, M. Laubenstein, Appl. Rad. Isot. , 191 (2000).28. L. Baudis et al. , JINST P08010, 201129. Geant4 Collaboration, S. Agostinelli et al. , Nucl. In-strum. Methods A , 250 (2003).30. Geant4 Collaboration, J. Allison et al. , IEEE Trans.Nucl. Sci. , 270 (2006).31. M. Boswell et al. , IEEE Trans. Nucl. Sci. et al. , Nucl. Instrum. Methods A606 , 790(2009).33. A.V. Dementyev and N.M. Sobolevsky, Radiat. Mea-sure. , 553 (1999).34. I. Barabanov et al. , Nucl. Instrum. Methods B251 , 115(2006).35. W.B. Wilson et al. , SOURCES-4A, Technical ReportLA-441 13539-MS Los ALamos (1999).36. M. Agostini et al. , Nucl. Physics B Proceedings Supple-ment , 489 (2012).37. Canberra Semiconductor, NV, Lammerdries 25, B-2250,Olen, Belgium. 38. M. Barnab´e Heider et al. , Procs. NSS08, IEEE Conf.publication 3513 (2008).39. M. Barnab´e Heider, PhD thesis, U. Heidelberg, May2009.40. H.V. Klapdor-Kleingrothaus, Nucl. Instrum. Methods
A566 , 472 (2006).41. J. Hellmig, H.V. KlapdorKleingrothaus, Z. Phys.
A359 ,351 (1997); L. Baudis et al. , Nucl. Instrum. Methods
A481 , 149 (2002).42. O. Chkvorets, PhD Thesis, U. Heidelberg, 2008.43. High Activity Disposal Experimental Site (HADES) ofthe Belgian Nuclear Research Center SCK · CEN, Boere-tang 200, BE-2400 Mol, Belgium.44. H.V. Klapdor-Kleingrothaus et al. , Nucl. Instrum.Methods
A522 , 371 (2004).45. V. Khournoukov, priv. communication; measurementsperformed by the Botchvar Insitute, Moscow, Russia.46. J. R. De Laeter et al. , Pure Appl. Chem. , 683 (2003).47. M. Barnab´e Heider et al. , JINST , P10007 (2010).48. Currently known as Joint Stock Company “ProductionAssociation Electrochemical Plant” (JSC “PA Electro-chemical Plant”), uranium enrichment enterprise of theState Atomic Energy Corporation “Rosatom”.49. PPM Pure Metals, GmbH, in Langelsheim, Germany.50. Canberra Oak Ridge, USA.51. E. Haller, Advances in Physics , 93 (1981).52. CANBERRA Broad Energy Ge (BEGe) Detector, .53. D. Budj´aˇs et al. , JINST , P10007 (2009).54. M. Agostini et al. , JINST , P03005 (2011).55. K. Zuber, PhD thesis, U. Heidelberg, 1992; analysis ofANG 1 by J. Echternach.56. F.T. Avignone, priv. comm., 2011.57. W. Maneschg et al. , Nucl. Instrum. Methods A593 , 448(2008).58. Torlon type 4503, Torlon is a trademark of Solvay Ad-vanced Polymers.59. G. Heusser et al. , Appl. Rad. Isot. , 691 (2000).60. Synthetic charcoal by Carbo-Act International (NL).61. Makrolon is a registered trade mark of BASF.62. Ch. Haberstroh, Advances in Cryogenic Engineering ,1201 (2008).63. LD301 HART Pressure Transmitter, SMAR ResearchCorporation, NY, USA.64. M. Balata, et al. , Nucl. Instrum. Methods A 370 , 605(1996); M.G. Giammarchi et al. , Ultrapure Water ,59 (1996).65. Institute of Environmental Science and Technology, Ar-lington Heights (IL), USA.66. Habia cable, Upplands V¨asby, Sweden.67. Sami S.p.a, Liscate (MI) Italy.68. Teledyne Reynolds UK, Newbury, Berkshire, UK.69. F. Froborg, PhD thesis, U. Zurich, 2012.70. National Instruments Corporation, Austin (TX), USA.71. M. Knapp, PhD thesis, Kepler Center for Astro andParticle Physics, U. T¨ubingen, 2009.72. Daylighting Film DF2000MA (VM2000), 3M, St. Paul,MN, USA.73. RG213C/U by JOWO Systemtechnik, Delmenhorst,Germany.74. CAEN S.p.A., Viareggio, Italy.75. F. Ritter et al. , Nucl. Instrum. Methods A617 (2010)420.76. F. Ritter, PhD thesis, Kepler Center for Astro and Par-ticle Physics, U. T¨ubingen, 2012.77. Scintillator UPS-923A is based on polystyrol with ad-ditive of PTP (2%) and POPOP (0.03%) built by theKharkov Crystal Institute, Ukraine.178. HM6780-2 (Ø 8 mm) Hamamatsu Photonics K.K.,Hamamatsu City, Japan.79. PMT-85 (Ø 30 mm) by Kvadrotech, Moscow, RU.80. S-type Y11(250)M, Kuraray, Japan. (The middle of thefiber is made of polystyrol with additive 0.02% of wave-length shifter Y11. The internal part of fiber consists ofPMMA, external from the fluorinated PMMA).81. S. Riboldi et al. , Procs. IEEE Nuclear Science Sympo-sium and Int. Workshop on Room Temperature Semi-conductor Detectors, Oct 30 - Nov 6, p. 1386 (2010).82. Polyflon Company, One Willard Road, Norwalk, CT06851, USA.83. Stycast, Emerson & Cuming, Westerlo, Belgium.84. R. Isocrate et al. , MD2S - Digital sampling electronicsfor the MARS detector , LNL Annual Report 2004, 226.85. C.A. Ur et al. , LNL Annual Report 2004, 228.86. J. Stein, F. Scheuer, W. Gast and A. Georgiev, Nucl.Instrum. Methods
B 113 , 141 (1996).87. Qt cross platform application frame work, Qt ProjectHosting, Oslo, Norway.88. STRUCK Innovative Systeme GmbH, 22399 Hamburg,Germany.89. PAS9817/AO DAC, Precision Analog Systems Co.,Plantation (FL) USA.90. M. Agostini et al. , J. Phys., Conf. Ser. , 042027(2012).91. Rene Brun and Fons Rademakers, ROOT Nucl. In-strum. Methods
A 389 , 81 (1997).92. M. Agostini, L. Pandola, P. Zavarise and O. Volynets,JINST , P08013 (2011).93. M. Agostini, L. Pandola and P. Zavarise, J. Phys.: Conf.Ser. , 012047 (2012).94. P. Zavarise et al. , J. Phys. Conf. Ser , 042028 (2012).95. Hewlett Packard ProCurve Switch 2810-48G.96. DELL Power Edge R300, Quad Core Xeon X3323, 2.5GHz.97. R. Brugnera et al. , JINST , P100712 (2012).98. PostgreSQL Database Management System.99. Ajax (Asynchronous JavaScript and XML), J.J. Garrett Ajax: A New Approach to Web Applications , AdaptivePath 2005.100. H. Neder, et al. , Appl. Rad. Isot. , 191 (2000).101. G. Heusser et al. , Procs. International Conference onIsotopes in Environmental Studies, Elsevier, Editors:P.P. Povinec and J.A. Sanchez-Cabeza, p. 495 (2006).102. R. Wink et al. , Nucl. Instrum. Methods A329 , 541(1993).103. D. Budj´aˇs et al. , Appl. Rad. Isot. , 755 (2009).104. G. Zuzel and H. Simgen, Appl. Rad. Isot. , 889 (2009).105. G. Heusser et al. , Appl. Rad. Isot. , 691 (2000).106. H. Simgen and G. Zuzel, Appl. Rad. Isot. , 922 (2009).107. Borexino collaboration; C. Arpesella et al. , Astropart.Phys. , 1 (2002).108. DuPont Performance Elastomers, Wilmington (DE),USA., or Ethylen-Propylen-Dien-Monomer (EPDM).109. Pfeiffer Vakuum, Asslar, Germany.110. BD Sensors, Thierstein, Germany.111. B. Bellini et al. , submitted to JCAP, arXiv:1202.6403112. G. Douysset et al. , Phys. Rev. Lett. , 4259 (2001).113. V.D. Ashitkov et al. , Instruments and ExperimentalTechniques, , 153-160 (2003).114. M. Laubenstein and G. Heusser, Appl. Rad. and Iso-topes , 750 (2009).115. A. Domula, PhD thesis in preparation, TU Dresden.116. G. Meierhofer et al. , Phys. Rev. C , 027603 (2010);G. Meierhofer et al. , Eur. Phys. J. A40 , 61 (2009). 117. J. Marganiec et al. , Phys. Rev. C , 065802 (2009); J.Marganiec et al. , J. Phys. G35 , 014022 (2008).118. G. Meierhofer et al. , Eur. Phys. J.
A48 , 20 (2012).119. P. Benetti et al. , Nucl. Instrum. Methods